Expression of O-Linked N-Acetylglucosamine Modified Proteins and Production and Characterization of Chlamydia Trachomatis CT663 Octavia Y

Louisiana State University LSU Digital Commons

LSU Doctoral Dissertations Graduate School

2014 Expression of O-linked N-acetylglucosamine modified proteins and production and characterization of chlamydia trachomatis CT663 Octavia Y. Goodwin Louisiana State University and Agricultural and Mechanical College, [email protected]

Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_dissertations Part of the Chemistry Commons

Recommended Citation Goodwin, Octavia Y., "Expression of O-linked N-acetylglucosamine modified proteins and production and characterization of chlamydia trachomatis CT663" (2014). LSU Doctoral Dissertations. 3325. https://digitalcommons.lsu.edu/gradschool_dissertations/3325

This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSU Doctoral Dissertations by an authorized graduate school editor of LSU Digital Commons. For more information, please [email protected]. EXPRESSION OF O-LINKED N-ACETYLGLUCOSAMINE MODIFIED PROTEINS AND PRODUCTION AND CHARACTERIZATION OF CHLAMYDIA TRACHOMATIS CT663

A Dissertation

Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College in partial fulfillment of the requirements for the degree of Doctor of Philosophy

The Department of Chemistry

by Octavia Yvonne Goodwin B. S., Georgia Southern University, 2008 December 2014

To my best friend and other half, Steven Calloway, for his patience, love, encouragement, and faith in me. To my mother, Judy Goodwin, for her unconditional love and support over the years. To my grandmother, Inez Walker, for always making my heart smile. To my brother, Jeffery Bolton, for always believing in me. To my entire family past, present, and future.

Acknowledgements I would like to thank my advisor, Dr. Macnaughtan, for guidance, support,

and faith in my abilities. I also would like to thank my committee members, Dr.

Isiah Warner, Dr. Samuel Gilman, and Dr. Aaron Smith for their guidance through

this process.

I would like to thank the Macnaughtan research group for their friendship

and support. Amid Paudyal, you have been so kind and helpful. Maggie

Thomasson, you are one of the best writers I know and I appreciate your

willingness to proofread. Dr. Kevin Roberson, although you are as stubborn as

an ox, you have been one of my best friends and I appreciate you. I would like to

say a special thanks to my “skinny new post doc”, Dr. Michelle Sweeney, you

have taught me so much about being professional and you have been an

exceptional friend.

I would also like to thank my undergraduate advisor, Dr. Karen Welch.

You believed in me when I did not believe in myself. I would have never been

brave enough to apply for graduate school without you. You were my first lab

mother and I will always love you.

Finally, I would not have made it through this process without my mother,

grandmother, sisters, and my best friend. Judy Goodwin, Inez Walker, Nikkie

Goodwin, Lakiesha Goodwin, Atiya Jordan, and Steven Calloway you are the

wind beneath my wings. I love you all to pieces.

iii

Table of Contents

Acknowledgements ...... iii

List of Tables ...... viii

List of Figures ...... ix

List of Abbreviations ...... xiv

Abstract ...... xv

Chapter 1: Background ...... 1 1. 1 Proteins ...... 1 1. 2 Protein Structure ...... 2 1. 3 Protein function as it relates to structure ...... 4 1. 4 Post translational modifications ...... 4 1. 5 O-GlcNAc in the hexosamine biosynthesis pathway...... 5 1. 6 O-GlcNAc is not your typical glycosylation ...... 6 1. 7 O-GlcNAc is a regulatory enzyme that is implicated in disease when it is dysregulated ...... 8 1. 8 Inhibitors of hexosaminidase ...... 10 1. 9 Methods for O-GlcNAc detection ...... 11 1. 9. 1 Galactosyltransferase-based methods ...... 11 1. 9. 2 Lectin blot ...... 11 1. 9. 3 Western blot ...... 12 1. 9. 4 Chemical Derivatization – BEMAD ...... 12 1. 10 Methods to study O-GlcNAc-modified proteins ...... 14 1. 11 Challenges ...... 14 1. 12 In vivo expression of O-GlcNAc-modified protein ...... 14 1. 13 Molecular cloning and gene expression ...... 15 1. 14 Recombinant protein expression and purification ...... 16 1. 15 Co-expression of target proteins with OGT ...... 17 1. 16 NMR ...... 18

Chapter 2: E. coli sabotages the in vivo production of O-linked-N- acetylglucosamine (O-GlcNAc) modified proteins ...... 21 2. 1 Introduction ...... 21 2. 2 Materials and methods ...... 22 2. 2. 1 Materials ...... 22 2. 2. 2 Cloning and expression vectors ...... 24 2. 2. 3 E. coli cultures and protein co-expression ...... 25 2. 2. 4 O-GlcNAc detection and quantification ...... 26 2. 2. 5 Partial purification of E. coli β-N-acetylglucosaminidase ...... 27 2. 2. 6 E. coli β-N-acetylglucosaminidase detection assays ...... 28 2. 2. 7 Kinetic assays ...... 29 2. 2. 8 Kinetic analysis ...... 30 2. 2. 9 Inhibition of E. coli β-N-acetylglucosaminidase activity ...... 31 2. 2. 10 NMR experiments to detect E. coli β-N-acetylglucosaminidase hydrolysis of an O-GlcNAc-modified peptide ...... 31 2. 3 Results and discussion ...... 33 2. 3. 1 Co-expression of OGT with substrate proteins using E. coli ...... 33 2. 3. 2 Inhibiting NagZ activity in E. coli increases in vivo O-GlcNAc modification ...... 37 2. 3. 3 E. coli NagZ cleaves O-GlcNAc from a peptide ...... 41 2. 3. 4 Identifying the E. coli β-N-acetylglucosaminidase as NagZ ...... 43 2. 4 Conclusions ...... 44 2. 5 Acknowledgments...... 46

Chapter 3: Is NagZ the saboteur? ...... 47 Chapter 3. 1 Introduction ...... 47 3. 1. 1 The Hexosamine biosynthesis pathway in E. coli ...... 47 3. 1. 2 The role of β-N-acetylglucosaminidase (NagZ) in E. coli ...... 48 3. 2 Materials and Methods ...... 50 3. 2. 1 Materials ...... 50 3. 2. 2 Knocking out the nagZ gene in the Rosetta(DE3) pLysS strain ...... 51 3. 2. 3 Making the ΔnagZ Rosetta(DE3) pLysS cells competent ...... 51

3. 2. 4 Transformation ...... 52 3. 2. 5 Expression conditions ...... 52 3. 2. 6 Ni-affinity purification conditions ...... 54 3. 2. 7 SDS-PAGE and western blot analysis ...... 54 3. 2. 8 WGA agarose affinity chromatography ...... 55 3. 2. 9 Coomassie protein concentration assay ...... 56 3. 2. 10 Supplementing ∆nagZ E. coli cultures with sugar ...... 56 3. 2. 11 Resolubilized O-GlcNAc-modified ABL2 ...... 57 3. 2. 12 OGT activity assay ...... 57 3. 3 Results ...... 57 3. 3. 1 Expression of O-GlcNAc-modified ABL2 in ΔnagZ E. coli ...... 57 3. 3. 2 ΔnagZ Rosetta(DE3)pLysS sugar trials ...... 59 3. 3. 3 Resolubilized ABL2 ...... 61 3. 3. 4 Optimized expression of O-GlcNAc-modified ABL2 in ∆nagZ ...... 63 3. 3. 5 Comparison of OGT activity produced from human OGT (hOGT) and codon-optimized OGT (coOGT) ...... 65 3. 4 Conclusion ...... 66

Chapter 4: Chlamydia trachomatis CT663: Structure and Ligand Interaction Model ...... 69 4. 1 Introduction ...... 69 4. 1. 1 Chlamydia trachomatis ...... 69 4. 1. 2 Type III secretion ...... 70 4. 1. 3 CT663/Scc1-CopN complex ...... 72 4. 1. 4 Research goal ...... 73 4. 2 Materials and Methods ...... 74 4. 2. 1 Materials ...... 74 4. 2. 2 Structure predictions ...... 75 4. 2. 3 CT663 expression and purification ...... 75 4. 2. 4 CT663 and Scc1 co-expression and co-purification: ...... 75 4. 2. 5 CopN expression and purification ...... 76 4. 2. 6 Separation of CT663 and Scc1 ...... 76 4. 2. 7 Gel electrophoresis analysis ...... 77

4. 2. 8 Gel filtration analysis ...... 77 4. 2. 9 SPR Analysis ...... 78 4. 2. 10 NMR buffer optimization for CT663 ...... 79 4. 2. 11 15N-CT663/Scc1 NMR analysis ...... 80 4. 3 Results ...... 81 4. 3. 1 Structural predictions of the Chlamydial trachomatis T3SS effectors- chaperone complex CT663:SCC1:CopN ...... 81 4. 3. 2 Expression and Purification ...... 82 4. 3. 3 Separation of CT663/Scc1 complex ...... 83 4. 3. 4 Buffer dependent gel filtration analysis ...... 85 4. 3. 5 Detergent assisted Gel filtration ...... 86 4. 3. 6 SPR analysis of Ct663/Scc1 binding to CopN ...... 88 4. 3. 7 The magical CT663 sample from never never land ...... 89 4. 3. 8 NMR buffer optimization for CT663 ...... 90 4. 3. 8 15N-CT663:Scc1 NMR analysis ...... 91 4. 3. 9 NMR analysis of the CT663/Scc1 complex ...... 92 4. 4 Conclusions ...... 93

References ...... 96

Appendix ...... 106 Letter of Permission ...... 106 Supplementary Figures...... 107

Vita ...... 112

vii

List of Tables Table 3-1 The recipe for 1 L of minimal growth media, 1 L of 10X M9 salts, and 100 mL of 1000X trace mineral ...... 53 Table 4-1 This table summarized the resulting oligomeric state of CT663 and Scc1/CT663 complex in different buffer conditions...... 87

viii

List of Figures

Figure 1-1 Synthesis and structure of the peptide backbone. http://www. molecularsciences. org/book/export/html/253 ...... 1 Figure 1-2 Transcription and translation of proteins. http://www. tokresource. org/tok_classes/biobiobio/biomenu/transcription_translation/ ...... 2 Figure 1-3 Protein structure. https://www. mun. ca/biology/scarr/iGen3_06-04. html ...... 4 Figure 1-4 The HBP pathway in humans. Yuzwa, S. A. and D. J. Vocadlo, O- glcnac and neurodegeneration: Biochemical mechanisms and potential roles in alzheimer's disease and beyond. Chemical Society Reviews, 2014. 43(19): p. 6839-6858...... 7 Figure 1- 5 Enzymatic cycling of O-GlcNAc by OGT and OGA. Hart, G. W. , M. P. Housley, and C. Slawson, Cycling of o-linked beta-n-acetylglucosamine on nucleocytoplasmic proteins. Nature, 2007. 446(7139): p. 1017-22...... 8 Figure 1-6 Methods of O-GlcNAc detection. Banerjee, P. S. , G. W. Hart, and J. W. Cho, Chemical approaches to study o-glcnacylation. Chemical Society Reviews, 2013. 42(10): p. 4345-4357 ...... 13 Figure 1-7 Steps involved in molecular cloning. http://cnx. org/contents/185cbf87- c72e-48f5-b51e-f14f21b5eabd@9. 45:88/Biology ...... 17 Figure 1-8 CBCANH experiment. http://www. protein-nmr. org. uk/pictures/experiment_types/cbcanh. pngc ...... 20

Figure 2-1The O-GlcNAc modification of CREB1 disrupts binding with TAFII130...... 22 Figure 2-2 The Ni-affinity purified soluble lysate and pellet from co-expression of CREB1-Q2 (pET-28b-CREB1-Q2) and OGT (pET-43. 1-OGT) in E. coli BL21- Gold(DE3) analyzed with (a) Coomassie-stained SDS-PAGE and (b) FITC-WGA lectin blot detecting O-GlcNAc. The bands corresponding to CREB1-Q2 are indicated with asterisks...... 34 Figure 2-3 The soluble lysate and pellet from co-expression of CREB1-Q2-bZIP (pET28b-CREB1-Q2-bZIP) and OGT (pET-43. 1-OGT) in E. coli BL21-Gold(DE3) analyzed with (a) Coomassie-stained SDS-PAGE and (b) FITC-WGA lectin blot detecting O-GlcNAc. The bands corresponding to CREB1-Q2-bZIP are indicated with asterisks...... 35 Figure 2-4 ABL2-FABD (pET-28b-ABL2-FABD) and OGT (pET-43. 1-OGT) were co-expressed in E. coli BL21-Gold(DE3) with increasing concentrations of PUGNAc in the growth media (0, 1, 2, and 10 µM PUGNAc). The soluble lysate (S) and pellet (P) from each culture was analyzed with (a) Coomassie-stained

SDS-PAGE and (b) anti-O-GlcNAc MAb CTD 110. 6 western blot. HEWL was used as a positive control in the last lane. (c) A plot of the relative band density of the western blot band to the Coomassie-stained SDS-PAGE gel band (indicated with an asterisk). For each band in (a) and (b), the background density was measured thrice, subtracted from each band, and used to estimate the propagated measurement error (represented as error bars in (c))...... 36 Figure 2-5 The initial velocities of E. coli β-N-acetylglucosaminidase, partially purified with CXC from BL21-Gold(DE3) cleared lysate, at varying MU-GlcNAc concentrations and with different concentrations of PUGNAc. The points represent the observed velocities, and the lines represent the best fits of the data to Eqs. (1) and (2)...... 39 Figure 2-6 The β-N-acetylglucosaminidase activity of cleared lysate with PUGNAc (0, 1, 2 µM) was measured using the pNP-GlcNAc spectrophotometric assay. The lysate was from the co-expression of ABL2-FABD (pET-28b-ABL2- FABD) and OGT (pET43. 1-OGT) in E. coli BL21-Gold(DE3). Each sample and control were run in triplicate, and the error bars represent the standard deviation of these measurements. The 405 nm absorbance values were converted to activity (right vertical axis) using the extinction coefficient of nitrophenolate...... 41 Figure 2-7 Fractions from CXC of BL21-Gold(DE3) lysate with β-N- acetylglucosaminidase activity toward MU-GlcNAc were pooled and tested for β- N-acetylglucosaminidase activity toward the synthetic O-GlcNAc-modified peptide, SVES(β-O-GlcNAc)GSADAK-NH2. (a) CXC fraction control, (b) CXC fraction control with 200 µM GlcNAc added, (c) CXC fraction with 200 µM glycopeptide, and (d) 200 µM glycopeptide control in CXC buffer. The spectra shown are expanded regions from 5. 1 to 5. 3 ppm, with a gray dashed line indicating the α-GlcNAc chemical shift (center of the doublet). All spectra were processed and expanded identically for direct comparison. The full spectra are included in Supplementary Figure 4. The glycopeptide control sample (d) has extra noise due to truncation artifacts from the water peak; the additional noise is small and does not affect spectral interpretation. Based on the 1H anomeric α- GlcNAc intensity in (b), approximately 40 µM GlcNAc was hydrolyzed by the E. coli β-N-acetylglucosaminidase in 5 h at 37 °C (c)...... 42 Figure 2-8 1H NMR spectra of (a) ∆nagZ K12 lysate, (b) lysate with 200 µM GlcNAc, and (c) lysate with 200 µM glycopeptide, SVES(β-O-GlcNAc)GSADAK- NH2. All samples were incubated at 37 °C for 22. 5 h to test for β-N- acetylglucosaminidase activity toward the synthetic O-GlcNAc-modified peptide. The spectra shown are expanded regions from 5. 1 to 5. 3 ppm, with a gray dashed line indicating the α-GlcNAc chemical shift (center of the doublet). All spectra were processed and expanded identically for direct comparison. The full spectra are included in Supplementary Figure 6. Based on the signal-to-noise

ratio of the 200 µM GlcNAc peak in (b), the limit of quantitation (LOQ) in (c) is 10 µM GlcNAc...... 44 Figure 3-1 Glucose, glucosamine, and N-acetylglucosamine are used by E. coli’s hexosamine biosynthesis pathway to make UDP-GlcNAc...... 48 Figure 3-2 NagZ cleaves the bond between GlcNAc and MurNAc, of the peptidoglycan layer [70]. Adapted from: Triassi, A. J. , M. S. Wheatley, M. Savka, H. M. Gan, R. C. J. Dobson, and A. O. B. Hudson, L,l-diaminopimelate aminotransferase (dapl): A putative target for the development of narrow- spectrum antibacterial compounds. Frontiers in Microbiology, 2014. 5...... 49 Figure 3-3 O-GlcNAc-modified ABL2 expressed in the ∆nagZ system was unstable and O-GlcNAc-ABL2 selectively precipitated. The soluble and insoluble samples both contain ABL2 as seen on the protein gel (a); however, the western blot (b) shows that O-GlcNAc-modified ABL2 seems to be exclusively in the pellet...... 58 Figure 3-4 Analysis of sugar supplemented expression of O-GlcNAc modified ABL2. The cleared lysate from cultures supplemented with glucose have a small amount of ABL2 (a) that is not glycosylated (b). The pellets corresponding to these samples contain about the same amount of ABL2 (c); however, the samples supplemented with glucose and glucosamine have an abundance of O- GlcNAc-modified ABL2 (d). The pellets containing PUGNAc express more ABL2 (c), but less O-GlcNAc-modified ABL2 compared to the glucose and glucosamine samples (d)...... 61 Figure 3-5 Purification of resolubilized O-GlcNAc-modified ABL2. The pellets were resolublized with 10% sarkosyl, diluted to 1%, and purified with nickel affinity chromatography. The results were very similar to previous data—there is not an abundance of ABL2 on the protein gel, (a) yet it is highly glycosylated (b)...... 63 Figure 3-6 Optimized expression of soluble O-GlcNAc-modified protein. A starter culture from a single colony was used to inoculate sugar-supplemented cultures in MM. The expression conditions were optimized to produce soluble O-GlcNAc- modified protein. The soluble lysate was purified and concentrated for SDS- PAGE and western blot analysis. The glucose sample was the only sample to produce a substantial amount of soluble O-GlcNAc-modified protein...... 64 Figure 3-7 Comparison of human OGT and codon optimized OGT. Anti-O- GlcNAc transferase was used to detect the amount of OGT present in the in vitro reaction. Anti-O-GlcNAc was used to detect the amount of O-GlcNAc-modified protein produced from the reaction...... 66 Figure 4-1 Developmental cycle of Chlamydia. Brunham, R. C. ; Rey-Ladino, J. , "Immunology of Chlamydia infection: Implications for a Chlamydia trachomatis vaccine," Nature Reviews Immunology 2005, 5 (2), 149-161...... 70

Figure 4-2 Diagram of animal pathogenic bacterial T3SS similar to that of Chlamydia [86]. Büttner D , and He S Y Type III Protein secretion in plant pathogenic bacteria. Plant Physiology. 2009;150:1656-1664...... 71 Figure 4-3 The CopN:CT663:Scc1 structure is modeled using Yersinia pestis virulence factor YopN in complex with the heterodimeric chaperone SycN-YscB by Phyre2...... 82 Figure 4-4 SDS-PAGE and native PAGE analysis of CT663, Scc1/CT663, and CopN. The protein molecular weight ladder is shown in lane 1 in the (a) SDS- PAGE and (b) native PAGE gel images. The molecular weight of CT663, Scc1/Ct663 and CopN are 14. 7 kDa, 30. 9 kDa, and 45. 2 kDa respectively. .. 83 Figure 4-5 CT663-Scc1 interaction. Purified CT663 and CT663:Scc1 complex are shown in lanes 2 and 3, respectively. The purified CT663-Scc1 complex was immobilized on 1 mL of Ni-affinity media via a 6XHis tag on Scc1 and washed with 20 mL of wash buffer with 6 M urea (lane 4), 8 M urea (lane 5), and 2% SDS (lane 6). Even under strong denaturing conditions of 8 M urea, some CT663 remained bound to 6XHis-tagged Scc1. SDS (2%) was sufficient to disrupt the CT663:Scc1 interaction, resulting in pure Scc1...... 84 Figure 4-6 Gel filtration analysis of CT663 and CT663:Scc1 at various buffer conditions. Two gel filtration column were used to analyze the samples: (A) a HiPrep Sephacryl S-100 HR column using a 1 mL sample volume and (B) an Enrich SEC 70 column using 50 . MolecularL sample weighvolumest standards were used to determine the linear relationship between the elution volume and the logarithm of the molecular weight for each column/buffer combination. The molecular weight of the samples were determined using the best-fit linear function based on regression analysis to determine the oligomeric state of the samples in each buffer. The sample concentrations were 81-152 µM except for the CT663:Scc1 complex in sodium acetate, which was 27 µM due to its instability in this buffer. Details about the sample conditions are described in Table 4-1. (A) In Tris/saline buffer at pH 8. 0, CT663 (solid black line) predominantly exists as higher order oligomers (>300 kDa) and as a homodimer while the CT663:Scc1 complex (dashed black line) predominantly forms a stable tetramer. In acetate buffer at pH 4. 5, CT663 and the CT663:Scc1 complex are unstable. CT663 readily precipitates, but the complex stability was better and showed that soluble CT663:Scc1 in acetate buffer (gray solid line) exists as a heterodimer. (B) In phosphate buffer at pH 7. 3 (solid black line), CT663 forms a homodimer and in Tris buffer at pH 7. 4 (dashed black line), CT663 forms homotrimers...... 85 Figure 4-7 Detergent effects on CT663 oligomerization. (A) CT663 samples were run in 20 mM Tris, pH 7. 4 buffer (solid black line) with 1 mM SDS (dotted black line), 10 mM SDS (dashed black line), and 0. 01% Triton X-100 (dash-dotted black line). The predicted elution volumes for CT663 oligomers (+) were

xii calculated using the calibration curve (Table 4. 1). The number of monomer units for each oligomer is shown above the marker. (B) CT663 samples were run in 50 mM potassium-sodium phosphate, pH 7. 3 buffer (solid gray line) with 0. 01% Triton X-100 (dotted gray line) and 1% PEG (dashed gray line). The predicted elution volumes for CT663 oligomers (+) are shown as in (A). Each sample was 50 -NaL ofphosphate, 81 M CT663 pH 7 .in 3 50buffer, mM exceptK for the sample run in 10 mM SDS/Tris buffer. This sample was diluted by a factor of 4 with the running buffer to avoid SDS-phosphate precipitation. The absorbance values for this sample were multiplied by 4 for direct comparison to the other samples...... 87 Figure 4-8 The SPR sensogram corresponding to the binding event between the Scc1/CT663 complex and CopN. The calculated Kd=3. 02 +/- 0. 17 nM...... 89 Figure 4-9 1H-15N HSQC of 15N labeled CT663. Each peak corresponds to an 1H- 15N bond from the backbone or a side chain of the protein...... 90 Figure 4-10 1H-15N HSQC of 15N labeled CT663 in 20 mM sodium phosphate, 20 mM tris, and 20 mM MES...... 91 Figure 4-11 Scc1 perturbs the amide peaks of 15N-CT663. An overlay of 1H-15N HSQC spectra of 15N-CT663 (red) and 15N-CT663:Scc1 (blue). The inset illustrates the selective shifts of the putative glycine amide peaks upon binding Scc1. Note: No peaks are covered by the inset...... 92 Figure 4-12 1H-15N HSQC of 15N labeled Scc1/CT663. The inset is a zoomed view to show the peak dispersion...... 93

xiii

List of Abbreviations O-GlcNAc O-linked beta-N-acetylglucosamine OGT O-GlcNAc transferase OGA O-GlcNAcase Glu glucose GlcN glucosamine GlcNAc N-acetylglucosamine UDP- Uridine diphosphate N- GlcNAc acetylglucosamine WGA Wheat germ agglutinin HBP Hexosamine biosynthesis pathway NagZ β-N-acetylglucosaminidase

xiv

Abstract O-linked β-N-acetylglucosamine is a regulatory post translational

modification. This modification occurs on nearly all functional classes of proteins, in the nucleus and cytoplasm. O-GlcNAc is added to serine or threonine by O-

GlcNAc transferase and removed by O-GlcNAcase. Previous attempts to study

O-GlcNAc-modified proteins have resulted in low yields, making 3-dimensional structure determination impossible. In this dissertation O-GlcNAc transferase will be co-expressed with domains of human cAMP responsive element-binding protein (CREB1) and Abelson tyrosine-kinase 2 (ABL2) in E. coli, to produce O-

GlcNAc-modified protein. The O-GlcNAc-modified protein was expressed in a variety of E. coli cell lines at a variety of conditions, but only small quantities of insoluble protein were produced. A glycosidase was suspected due to the disappearance of the O-GlcNAc modification from the protein. O-(2-acetamido-2- dexoy-dglucopyranosylidene) amino-N-phenylcarbamate (PUGNAc), a β-N- acetylglucosaminidase inhibitor, was added to the culture media and increased the production of O-GlcNAc-modified protein. This was the first evidence that β-

N-acetylglucosaminidase (NagZ), an E. coli enzyme, cleaves O-GlcNAc from proteins in vivo. NagZ was isolated and shown to cleave O-GlcNAc from a synthetic O-GlcNAc-peptide in vitro. In E. coli, NagZ cleaves the GlcNAc-β1,4-

N-acetylmuramic acid linkage to recycle peptidoglycan in the cytoplasm. A

NagZ knockout showed no activity towards the O-GlcNAc-peptide, confirming

NagZ as the enzyme responsible for cleaving O-GlcNAc from our glycoprotein expressed in vivo. O-GlcNAc-modified protein produced by the NagZ knockout

(∆NagZ) co-expression system is highly glycosylated and can be resolubilized

from the pellet. The ∆NagZ is a step closer to production of milligram quantities of O-GlcNAc-modified protein for structure determination.

xvi

Chapter 1: Background

1. 1 Proteins

Proteins are macromolecules that are present in all living systems. They which play essential roles in many cellular processes such as transport, storage, cellular signaling and immune response [1]. A major class of proteins, known as enzymes, act as catalyst for many chemical reactions in biological systems. They are composed of 20 amino acid building blocks. These amino acids, containing a carboxyl end, amino end and side chain, are linked together when a dehydration reaction occurs between two amino acids and forms a peptide bond (Figure 1-1).

Figure 1-1 Synthesis and structure of the peptide backbone. http://www. molecularsciences. org/book/export/html/253

Proteins are synthesized through two processes called transcription and translation. During transcription, RNA polymerase transcribes single strands of DNA into complimentary strands of RNA called messenger RNA (mRNA). The mRNA is then translated into a polypeptide chain by ribosomes, composed of ribosomal proteins,

1 ribosomal RNA (rRNA), and transfer RNA (tRNA). tRNA is unlike mRNA in that it carries

amino acids. One side of tRNA has nucleotide base pairs and the other has the amino acids that correspond to the nucleotide base pairs. The ribosome synthesizes a protein by reading mRNA, binding the appropriate tRNA, and creating peptide bonds between each amino acid (Figure1-2) [2].

Figure 1-2 Transcription and translation of proteins. http://www. tokresource. org/tok_classes/biobiobio/biomenu/transcription_translation/

1. 2 Protein Structure

The overall 3-dimensional (3D) structure or tertiary structure of a protein is

dependent upon its primary and secondary structure. The primary structure of a protein

is its amino acid sequence. The secondary structure describes how this chain of amino

acids behave in relation to their bond neighbor. Secondary structure consists of alpha

2 helices, beta sheets and non-structured regions called loops. Alpha helices form a

spiral-like shape with 3. 6 amino acids for every turn. The protein backbone forms the

inside of the coil with the amino acid side chains at the exterior. The spiral or coil shape

is stabilized by intramolecular hydrogen bonds formed between the amide group hydrogen and the oxygen of a carboxyl group 4 residues away (Figure 1-3). Beta sheets

are flat or linear in shape and are also held together by hydrogen bonds. These bonds

differ from those in the alpha helix in that they occur between separate chains. Also, the

intramolecular hydrogen bond is between the hydrogen of an amide on one chain and

the oxygen of a carboxyl of a neighboring chain. The alpha helices, beta sheet and

loops fold to make subunits of the proteins. These subunits describe a protein’s 3D

tertiary structure. Finally, these tertiary structures of protein monomers fit together to

make protein complexes or multimers. The structures resulting from these multimers is

quaternary structure, which is the overall folded state of the entire complex (Figure1-3).

The tertiary and quaternary structures are held together by hydrogen bonds, ionic

bonds, hydrophobic interactions and disulfide bonds between side chains [3].

3 Figure 1-3 Protein structure. https://www. mun. ca/biology/scarr/iGen3_06-04. html

1. 3 Protein function as it relates to structure

Protein function depends on protein structure. In most cases, proteins have to be folded in a certain conformation to carry out their function. For example, in lock and key ligand binding, enzymes show specificity for substrates by forming an active cite based on the enzyme’s structure. This substrate binding event is often part of a cascade of events required for function to take place [3].

1. 4 Post translational modifications

A post translation modification (PTM) is a covalent chemical modification that occurs after the protein has been translated. There are many types of PTMs, such as methylation, ubiquitination, phosphorylation, and glycosylation. Methylation adds a

4 methyl group to a lysine or arginine. Ubiquitination is the addition of an entire ubiquitin

protein molecule to a lysine residue. Phosphorylation and glycosylation are the two

most common PTMs. Phosphorylation is the addition of a phosphate at a serine and/or

threonine residue. Glycosylation can be N-linked and O-linked. N-linked glycosylation normally consists of the addition of a polysaccharide chain at an asparagine, as long as the asparagine is part of a tri-peptide motif consisting of the asparagine, any amino acid and then a serine or threonine (N, X, S/T). This consensus sequence is a tri-peptide within the sequence of the protein where the N-linked glycosylation may occur.

Knowledge of this consensus sequence allows us to predict possible glycosylation sites.

Most O-linked glycosylations are similar to N-linked glycosylations, in that they are long chain polysaccharides. The exception is O-linked N-acetylglucosamine (O-GlcNAc), which is the addition of a monosaccharide instead of a polysaccharide [4].

1. 5 O-GlcNAc in the hexosamine biosynthesis pathway

The hexosamine biosynthesis pathway (HBP) is a metabolic pathway that produces uridine diphosphate N-acetylglucosamine (UDP-GlcNAc), the substrate for O-

GlcNAcylation, and other nucleotide hexosamines (Figure 1-4). Hexosamines, also called amino sugars, are sugars that have an amino group (-NH2) in place of a hydroxyl

group. The HBP varies based on the organism. In humans, glucose enters the cell and

is phosphorylated by hexokinase. Only 2-5% of glucose that enters the cell enters the

HBP, therefore the level of O-GlcNAcylation is dependent on nutrient availability [5].

Once glucose is phosphorylated and converted to fructose-6-phosphate, it can enter the

HBP or the main glycolytic/glycogen pathway. The HBP is initiated by

glutamine/fructose aminotransferase (GFAT). GFAT is an enzyme that transfers an

5 amino group to fructose-6-phophate from the amino acid glutamine, producing glucosamine-6-phosphate [6]. It is important to add that glucosamine-6-phophate can also be produced by glucosamine that enters the cell and is phosphorylated by hexokinase [7]. EMeg32, a glucosamine-6-phospate N-acetyltransferase, in an enzyme that acetylates glucosamine-6-phosphate using acetyl-CoA as a substrate [8]. N- acetylglucosamine-6-phosphate, produced by the acetylation reaction, is interconverted to N-acetylglucosamine-1-phosphate by N-acetyl phospho-glucosamine mutase

(AGM1). N-acetylglucosamine-1-phosphate is finally uridylated to UDP-GlcNAc by UDP-

GlcNAc pyrophosphorylase (UAP1) [9].

1. 6 O-GlcNAc is not your typical glycosylation

O-GlcNAc is an O-linked glycosylation that occurs at a serine or a threonine. O-GlcNAc is cycled on and off of protein substrates by enzymes. It is added by the enzyme O-

GlcNAc transferase (OGT) and remove by O-GlcNAcase (OGA) (Figure 1-5) [10]. This

O-linked modification is unlike most glycosylation. Only one sugar is added, as opposed to a longer polysaccharide chain. There is no consensus sequence recognized by OGT, which makes O-GlcNAc-site prediction difficult [11]. O-GlcNAc has a reciprocal relationship with phosphorylation. They can occur on the same protein at the same time or at a specific residue interchangeably [12]. This relationship with phosphorylation is used by most prediction algorithms to identify possible O-GlcNAc modification sites [12].

In mammals, most protein glycosylation occurs in the extracellular matrix or the secretion pathway. However, O-GlcNAc can be found in the cytoplasm, nucleus and mitochondria.

Figure 1-4 The HBP pathway in humans. Yuzwa, S. A. and D. J. Vocadlo, O-glcnac and neurodegeneration: Biochemical mechanisms and potential roles in alzheimer's disease and beyond. Chemical Society Reviews, 2014. 43(19): p. 6839-6858.

Figure 1- 5 Enzymatic cycling of O-GlcNAc by OGT and OGA. Hart, G. W. , M. P. Housley, and C. Slawson, Cycling of o-linked beta-n-acetylglucosamine on nucleocytoplasmic proteins. Nature, 2007. 446(7139): p. 1017-22.

1. 7 O-GlcNAc is a regulatory enzyme that is implicated in disease when it is dysregulated

The O-GlcNAc modification was first discovered by Gerald Hart in 1984 and is

dynamic and regulatory in nature [13]. O-GlcNAc-modified proteins regulate many

biological processes, such as transcription, translation, protein degradation, cell

signaling, cell trafficking, apoptosis and cell cycle control [14]. It also regulates some

physiological events, such as the circadian rhythm, memory formation, and learning

[15]. When the OGT gene is knocked out in mice, it is embryonic lethal, therefore this

regulatory modification is essential for life [16].

Cyclic AMP-response element binding protein 1 (CREB1) is a transcription factor

that binds to TATA-associated factor (TafII130), a coactivator, to induce transcription of the cyclic AMP response element (CRE) gene [17]. When O-GlcNAc is bound to

CREB1, association with TafII130 is inhibited and transcription is repressed [17].

Repression of the CRE gene leads to reduced growth in axons and dendrites in mice

brain cells [18]. Axon and dendrite growth and connection is how memories are formed.

8 Dysregulated O-GlcNAcylation of CREB1 leads to memory loss in mice and has been linked to Alzheimer’s disease [20].

In the brain, neurons (brain cells) consist of a cell body, dendrites, and axons.

The neuronal cell body is very similar to other cells and contains the nucleus, cytoplasm and cellular membrane. Dendrites carry information into the cell body through electric signals, while axons take information away from the cell body. Dendrites are connected

to the cell body at one end while the axons are at the opposite end of the cell body [21].

Microtubules are the major structural component of axons. They play a key role in

intracellular transport by carrying nutrients and other cellular cargo [21]. They are hollow

and tubular in shape and are composed of polar filamentous polymers of α/β tubulin

heterodimers. These microtubules are held together by MT binding proteins such as

tau. Hyperphosphorylated tau separates tau from the microtubules, causing them to fall

apart [22]. The hyperphosphorylated tau molecules aggregate within the neuron,

causing tangles, inhibiting neurotransmission and destroying the cell [23]. The O-

GlcNAc modification can compete with phosphorylation sites on tau. When OGA is

inhibited in rodents, the O-GlcNAc modification increases and the phosphorylation

decreases at pathologically relevant sites on Tau [23]. This increase in O-GlcNAcylation

blocks hyperphosphorylation and hinders aggregation [24].

O-GlcNAc also plays a role in diabetes and cancer. Recent research has

suggested that an increase of the O-GlcNAc modification in prostate cancer cells

promotes the progression of tumor cells. This elevation of O-GlcNAc is associated with

an increase in overall patient expiration [25]. Elevated O-GlcNAc levels have also been

shown to enhance migration/invasion and metastasis in breast cancer cells [26]. Over

9 expression of OGT, the enzyme that adds O-GlcNAc; and GFAT, the rate limiting step of the HBP; in various tissues leads to insulin resistance in mice [27]. Insulin is a hormone that carries glucose from the bloodstream into cells. Insulin resistance refers to the decreased efficiency of insulin for glucose uptake, metabolism and storage.

Insulin resistance is a major feature and cause of Type 2 diabetes [28].

1. 8 Inhibitors of hexosaminidase

O-GlcNAcase is a hexosaminidase that removes the O-GlcNAc modification from substrate proteins. Hexosaminidases are enzymes that remove hexosamines. There are several different hexosaminidases. When analyzing the O-GlcNAc modification in cells that contain a hexosaminidase, it is often necessary to inhibit the hexosaminidase in order to increase glycosylation and analyze the effect of hyperglycosylation.

Streptozotocin (STZ) was one of the first OGA inhibitors used [29]. Recent research suggest that STZ acts as a chemical poison rather than an OGA inhibitor [5]. O-(2-

Acetamido-2-deoxy-D-glucopyranosylidenamino) amino-N-phenylcarbamate (PUGNAc) is another inhibitor of OGA with a Ki value of 50 nM for OGT purified from rats [5].

Although PUGNAc inhibits OGA, it is not specific to OGA and inhibits other glycosidases [30]. PUGNAc inhibits lysosomal hexosaminidase with a Ki of 30-40 nM

[5]. Thiamet-G has a Ki of 21 nM for OGA and is a stable bioactive inhibitor, leading to increased tau glycosylation in mice brain [31]. In an effort to create a strong and selective inhibitor, Dorfmuller and coworkers developed a class of small molecule OGA inhibitors called GlcNAcstatins [32]. This class of inhibitors is more selective towards human OGA and increases O-GlcNAcylation by 2 to 3 fold [5].

10 1. 9 Methods for O-GlcNAc detection

1. 9. 1 Galactosyltransferase-based methods

Β-1,4-Galactosyltransferase is an enzyme that transfers a single galactose from

UDP-galactose to a terminal GlcNAc molecule (Figure 1-6b). This enzymatic labeling method was used in 1984 by Gerald Hart when he first discovered the O-GlcNAc modification [13]. Radioactive 3H-galactose was added to O-GlcNAcylated proteins and

the radioactivity was detected. This technique can be used to transfer a galactose with a

chemically reactive group to the O-GlcNAc residue [33]. This allows for biotin labeling,

which can then be used for chemiluminescent and western blot detection, or enrichment

for further analysis.

1. 9. 2 Lectin blot

A lectin blot is a technique in which a lectin is used to detect a glycoprotein

(Figure 1-6a). The lectin, which is a protein that recognizes and binds a specific

carbohydrate and binds carbohydrate based molecules, is conjugated to a fluorescent

tag for detection. Lectins are found in most animals, plants and microorganisms. Wheat

germ agglutinin (WGA) is the lectin that recognizes and binds to terminal GlcNAc or N-

acetylneuraminic acid residues [34]. WGA is bound to agarose beads and used as a

means of affinity purification in order to enrich the glycoprotein for further analysis.

Although WGA binds O-GlcNAc, it is not always selective. For this reason, other means

of detection are sometimes necessary.

11 1. 9. 3 Western blot

Western blots are very similar to lectin blots. Instead of using a lectin to bind and detect the protein of interest, primary and secondary antibodies are used (Figure 1-6a).

Proteins are transferred from a sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel onto a nitrocellulose or PVDF membrane. The

membrane is then blocked. The term blocked refers to the blocking of free space on the

membrane to insure proper antibody binding. The membrane is probed with the primary

antibody. This antibody binds to the epitope of interest. CTD 110. 8 and RL-2 are

common antibodies for the detection of O-GlcNAc. CTD 110. 8 binds specifically to O-

GlcNAc while RL2 less specific. A secondary antibody, which is conjugated to a

fluorescent tag, binds the primary antibody. Although this method of detection is very

similar to a lectin blot, it is more selective.

1. 9. 4 Chemical Derivatization – BEMAD

Mass spectrometry (MS) techniques that use high energy ionization methods

tend to fragment the O-GlcNAc modification before it can be analyzed. In 2002 Lance

Wells applied the beta elimination Michael addition (BEMAD) reaction, a two-step

organic reaction, to O-GlcNAc-modified proteins [35]. The BEMAD reaction replaces the

O-GlcNAc modification with another compound that is stable during MS analysis. The

first step of the reaction is the beta elimination of O-GlcNAc, followed by the addition of

a new compound via Michael addition of a nucleophile. Dithiothreitol (DTT) or biotin

pentylamine (BAP) are the nucleophiles of that have been used [36]. At the time, this

reaction was notable because it allowed detection of O-GlcNAc-modified protein and

site mapping. Site mapping is a process where the exact amino acid modified is

12 identified. Now, there are lower energy fragmentation techniques that allow the O-

GlcNAc modification to stay intact during site mapping [35].

Figure 1-6 Methods of O-GlcNAc detection. Banerjee, P. S. , G. W. Hart, and J. W. Cho, Chemical approaches to study o-glcnacylation. Chemical Society Reviews, 2013. 42(10): p. 4345-4357

13 1. 10 Methods to study O-GlcNAc-modified proteins

There have been numerous strategies to study O-GlcNAc-modified proteins. O-

GlcNAc-modified proteins have been extracted from animal tissues. OGT has been extracted and used for in vitro glycosylation. OGT has been expressed in E. coli and

used for in vitro glycosylation. Target proteins have been expressed in human cells that

have the OGT gene. The aim of methods is to identify O-GlcNAc-modified proteins.

WGA affinity chromatography is often used to enrich protein samples. The sites of

glycosylation are then mapped using MS. Mice and fly have been used to study to

effects of the O-GlcNAc modification on cellular processes related to a specific protein

[37, 38].

1. 11 Challenges

Although there have been major discoveries about the O-GlcNAc modification

since it was first discovered, there are still many limitations. The greatest of these is the sample size. Proteins extracted from animal tissue are low in concentration. The same is true for modified proteins expressed in human cells. In vitro reactions are

normally done on a small scale due to the low concentration of extracted OGT. Human

OGT expressed in E. coli has also been used for in vitro glycosylation reactions. Limited

sample concentration has made structural analysis of O-GlcNAc-modified proteins

impossible. NMR and X-ray crystallography require milligram quantities of protein when

the methods to date yield only micrograms of O-GlcNAc-modified protein.

1. 12 In vivo expression of O-GlcNAc-modified protein

Dysregulation of the O-GlcNAc modification, whether too much or too little, has

been implicated in disease. Given that the modification is essential for life and has been

14 linked to several diseases, understanding the effect of the modification on protein

structure is imperative. In this thesis, O-GlcNAc-modified proteins will be expressed in vivo. In this particular method, OGT and a target protein are co-expressed in E. coli.

This method is especially interesting because it enables the expression of large

amounts of modified target proteins. More importantly, this method allows isotopic

labeling for NMR structural analysis. E. coli was chosen over mammalian or insect cells

because it does not carry the OGT gene. E. coli is the most common expression system

for milligram quantities of protein for NMR and X-ray crystallography structural analysis

[39]. This method requires molecular cloning of the target proteins and OGT DNA into

expression vectors, transformation of the vectors into expression cells, protein

expression and purification. Once the protein has been successfully produced with and

without the modification, their 3D structures will be determined and compared. E. coli

requires sources of carbon and nitrogen to synthesize proteins. Growing E. coli with

15N-ammonium chloride and 13C-glucose produces proteins with isotopically labeled

backbones and side chains. This method also allows the isotopic labeling of the O-

GlcNAc modification.

1. 13 Molecular cloning and gene expression

Molecular cloning is a process where DNA encoding a target protein is inserted into a cloning vector. This process is usually achieved through polymerase chain reaction (PCR), restriction digestion, and ligation (Figure 1-7). PCR is an experiment

where a small fragment of DNA is amplified, creating many copies of that fragment.

PCR consist of 3 steps. First, the DNA is denatured separating it into single strands.

Primers anneal, or bind, to their consensus sequences at the 3’ end, and elongation

15 occurs. During elongation, RNA polymerase uses DNA base pair components to extend the primers, which makes a new double stranded DNA fragment. Every time this process occurs, the DNA fragments double. Once the fragment has been amplified, it is then digested with restriction enzymes. Restriction enzymes recognize specific DNA sequences and cleave them, leaving sticky or overlapping ends. This same digestion is done on the cloning vector. In the ligation step, the sticky ends are bound together by a ligase. The two pieces of DNA come together to form a plasmid, circular DNA, which is then transformed into competent E. coli cells. These cells make several copies of the plasmid that is used for protein expression.

1. 14 Recombinant protein expression and purification

Recombinant protein expression is a process in which DNA corresponding to a target protein is expressed from a vector. The expression vector is made of DNA that has all the necessary promoters and operators for protein expression in a bacterial host.

The cloned and recombined DNA is transformed into E. coli and isopropyl β-D-1- thiogalactopyranoside (IPTG) is used to induce expression of the protein. The cells are harvested and lysed through mechanical or chemical force. Soluble proteins are separated from the cell debris by centrifugation. The protein of interest is isolated through a purification process such as affinity chromatography. Affinity chromatography is a process were a protein is tagged with the addition of some sequence of amino acids that has a moderate binding affinity for a compound immobilized on some type of resin.

For example, nickel affinity chromatography uses the affinity between 6 histidine residues on the protein and Ni2+ bound to the resin to select for the protein of interest.

Figure 1-7 Steps involved in molecular cloning. http://cnx. org/contents/185cbf87-c72e- 48f5-b51e-f14f21b5eabd@9. 45:88/Biology

1. 15 Co-expression of target proteins with OGT

O-GlcNAc transferase (OGT) is the enzyme that adds O-GlcNAc to a target protein. Scientists began to question if a co-expression of OGT and a target protein would produce O-GlcNAc-modified protein. Although the mechanism of OGT is not fully

17 understood, we know that UDP-GlcNAc is the substrate for the addition of the O-

GlcNAc modification. E. coli produces UDP-GlcNAc within the cells, making the expression of O-GlcNAc-modified proteins possible in a co-expression system. David

Vocadlo has expressed a few O-GlcNAc-modified proteins using this method [24].

Expressing the modified protein in E. coli permits 15N and 13C isotopic labeling of the

protein, as well as the modification, for NMR structural analysis.

1. 16 NMR

NMR is an analytical technique where the nuclear spins of a molecule are

aligned by a constant magnetic field. An external oscillating magnetic field generated by

a radiofrequency pulse is used to transfer a population of nuclei into a higher energy

state with phase coherence. As the nuclei relax back to their lower energy state, the

oscillating magnetic field generated by the coherent spins is detected. To date, there is

no protein structure with the O-GlcNAc modification available by NMR or X-ray

crystallography. One of the major setbacks of X-ray crystallography is that the technique depends on the protein’s ability to be crystallized. It also only gives a static view of a protein’s structure. NMR allows for the detection of conformational changes under physiological conditions. Although NMR is limited by the size of proteins that can be analyzed (about 30 kDa), it is not a problem for this research because all target proteins are within the applicable size range.

Methods for determination of protein structure by NMR have become standard,

starting with backbone and side chain peak assignment, followed by distance and

restraint measurements, and concluding with the final calculation of the structure. 1H-

15N heteronuclear single quantum coherence spectroscopy (HSQC) is the first NMR

18 experiment performed on the purified isotopically labeled protein. This experiment gives

one peak for each amide group from each amino acid in the protein. 2D experiments

show the correlation between two nuclei, in this case the 1H attached to the 15N of the

amide group. Side chains containing amines also appear in 1H-15N HSQC experiments

as well. The dispersion, intensity and number of peaks gives valuable information about

the protein sample. Spectra with narrow dispersion between 7-8 ppm indicate that the

protein is unfolded or aggregated. Peak intensity is related to the concentration of the

sample and the oligomerization state. Optimal sample conditions yield greater than 95%

of expected backbone peaks. Once the conditions are optimized, additional experiments

are done for further assignments. The next experiment collected is CBCANH. This 3D

experiment shows which Cα’s and Cβ’s are connected to each amide group previously

identified from the 1H -15N HSQC experiment (Figure 1-8). In most cases, there are two

Cα’s and two Cβ’s peaks for each amide, the Cα/Cβ pair on the previous residue and

the Cα/Cβ pair directly connected to the amide. These Cα/Cβ pairs are indicated by the

residue number i and i -1. Because the i-1 Cα/Cβ pair is not always visible, an additional

experiment is done. CBCA(CO)NH only shows peaks corresponding to Cα/Cβ i-1.

These experiments allow scientists to identify the amino acid backbone peaks of the

protein. Chemical shifts give information about the identity of the side chains and further

experiments can be done if required.

The next set of experiments performed gives information about the secondary

structure by measuring scalar coupling (J-coupling). J-coupling is a through bond measurement that gives information about bond angles. Nuclear overhauser effects

(NOEs) are through space measurements and give information about protein

19 conformation. NOEs between neighboring residues also give secondary structural

information. 2D Nuclear overhauser spectroscopy (NOESY) is used to measure

distances between protons in close proximity. 15N-NOESY experiments give NOE

measurements for all hydrogens within 5 Å of an NH group. 13C-NOESY measures all the hydrogens within 5 Å of a CH group. Once the NOESY experiments are completed, there are automated programs that assign the peaks. These assignments, along with the previous assignments and distance restraints, are used to calculate a protein structure [40].

Figure 1-8 CBCANH experiment. http://www. protein-nmr. org. uk/pictures/experiment_types/cbcanh. pngc

20 Chapter 2: E. coli sabotages the in vivo production of O-linked-N- acetylglucosamine (O-GlcNAc) modified proteins

2. 1 Introduction

cAMP response element binding protein (CREB1) is a transpiration factor, that

plays a major role in the nervous system and brain. CREB1 is present in both neurons

and glial cells, which are the major components of the nervous system. CREB1 is well

known for its essential role in long-term memory and learning [41]. CREB1 is a human

protein that contains 341 amino acids divided into 4 domains. It includes 2 hydrophobic

and glutamine rich domains: Q1 at the N-terminus and Q2 in the middle of the

sequence. A kinase-inducible domain, KID, separates the Q1 and Q2 domains. The

bZip, basic leucine zipper, domain is at the C-terminus. CREB1 binds to DNA and

activates expression of proteins that contribute to synaptic plasticity [42]. Synaptic

plasticity refers to the process by which synapses are modified in structure and function

in response to different stimuli or environmental cues. Activity-dependent synaptic

plasticity is thought to be the underlying process of learning and memory [43]. Josselyn

et al. showed the first direct evidence that increasing CREB1 levels in a defined

mammalian brain region enhances long-term memory formation [44].

CREB1 is O-GlcNAcylated in vivo. The modification occurs at two residues

between amino acids 256-261 in the Q2 domain at S1er-260, Thr-256, Thr-259, or Thr-

261 [41]. Transcriptional activity is repressed when CREB1 is modified with O-GlcNAc because O-GlcNAc it inhibits CREB1’s association with the coactivator, TATA-

1 This chapter previously appeared as: Octavia Y. Goodwin, Maggie S. Thomasson, Aaron J. Lin, Michelle M. Sweeney, Megan A. Macnaughtan. E. coli sabotage the in vivo production of O-linked-β-N- acetylglucosamine-modified proteins. Journal of Biotechnology 168 (2013) 315-323.

21 associated factor (TAFII130) [41].

Figure 2-1The O-GlcNAc modification of CREB1 disrupts binding with TAFII130. Recent studies have shown that inhibition of O-GlcNAcylation leads to enhanced axonal

and dendritic growth and long-term memory [18]. A study done in mice showed that

mice expressing O-GlcNAcylated CREB1 had a decrease in long-term memory

compared to mice that expressed non-glycosylated CREB1 . In this chapter I describe

how CREB1 was cloned, expressed, and glycosylated in vivo for structural analysis.

2. 2 Materials and methods

2. 2. 1 Materials

Ovalbumin, hen egg white lysozyme, N-acetylglucosaminidase from Canavalia

ensiformis (Jackbean), fluorescein-conjugated wheat-germ agglutinin (FITC-WGA), glycerol, maleic acid, O-(2-acetamido-2-dexoyd-glucopyranosylidene)amino-N- phenylcarbamate (PUGNAc), 4,4-dimethyl-4-silapentane-1-sulfonic acid (DSS), sodium chloride (NaCl), sodium phosphate dibasic, sodium phosphate monobasic, bromophenol blue, dimethylformamide, 4-methylumbelliferyl-N-acetylglucosaminide

(MU-GlcNAc), sodium carbonate, p-nitrophenyl-β-N-acetylglucosamine (pNP-GlcNAc),

Coomassie brilliant blue R250, sodium dodecyl sulfate (SDS), glycine,

22 tris(hydroxymethyl)aminomethane (tris), 2-mercaptoethanol, His-Select Ni-affinity gel, imidazole, urea, acetic acid, methanol, and kanamycin were purchased from Sigma-

Aldrich. pET-28b vector, Rosetta(DE3) cells, Tuner(DE3) cells, BugBuster protein extraction reagent, and ampicillin were purchased from EMD Millipore/Novagen. A UNO

S1 cation exchange column, Precision Plus Protein Unstained Standards, 4–20% mini-

PROTEAN TGX precast gels, immun-blot low fluorescence PVDF membrane, and

Laemmli buffer were purchased from Bio-Rad. 2-log DNA ladder, 10-beta cells, Taq 5X

master mix, T4 DNA ligase, NdeI restriction enzyme, and XhoI restriction enzyme were

purchased from New England Biolabs. BioReagents EZ-Run Rec Protein Ladder, anti-

O-GlcNAc MAb CTD 110. 6, and DyLight 550-conjugate Goat anti-Mouse IgG

secondary antibody were purchased from Thermo Scientific/Pierce Antibodies. Luria

broth (LB) and agar were purchased from BD Biosciences. Artic Express(DE3) cells and

BL21-Gold(DE3) cells were purchased from Agilent Technologies, Strategene Products

Division. OverExpress C41(DE3) cells and OverExpress C43(DE3) cells were

purchased from Lucigen. Other chemicals and materials used include a glycopeptide

(SVES(β-O-GlcNAc)GSADAK-NH2, Sussex Research), deuterated water (D2O,

Cambridge Isotope Laboratories), isopropyl-D-1-thiogalactopyranoside (IPTG,

Affymetrix Anatrace), primers (Integrated DNA Technologies), a TSKgel G3000 SWxl

gel filtration column (Tosoh Bioscience LLC), full-length cDNA sequence of human

cAMP responsive element-binding protein in pCMV-Sport 6 vector (CREB1, Invitrogen),

QIAprep spin miniprep kit (Qiagen), Costar 96 well UV transparent flat bottom plates, 4

mm susceptibility matched NMR tube (D2O-matched, Shigemi Inc. ), K-12 nagZ

knockout strain (JW1093-1) and K-12 parent strain (BW25113) (Keio Collection, E. coli

23 Genetic Resources at Yale, CGSC, The Coli Genetic Stock Center) [45], and clone

HsCD00331895 of the F-actin binding domain of Abelson tyrosine-kinase 2 (ABL2-

FABD) in pET15 vector (DNASU Plasmid Repository, PSI:Biology-Materials Repository)

[46]. All water used was supplied from a Millipore Direct-Q 3 ultrapure water system.

2. 2. 2 Cloning and expression vectors

The human O-GlcNAc transferase (OGT) was expressed using the pET-43. 1

Ek/LIC vector with the full-length sequence provided by Dr. John Hanover [47]. The pET-43. 1 Ek/LIC-OGT includes an N-terminal NusA-, 6XHis-, and S-tags. The sequence of three substrate protein domains were cloned into the expression vector, pET-28b, with an N-terminal 6XHis tag: the second glutamine-rich domain, Q2, of human cAMP responsive element-binding protein (CREB1-Q2, amino acids 160–283), the Q2 and bZIP domains of CREB1 (CREB1-Q2-bZIP, amino acids 160–341), and the

F-actin binding domain of Abelson tyrosine-kinase 2 (ABL2-FABD). CREB1-Q2 and

CREB1-Q2-bZIP were cloned from cDNA of the full-length sequence using PCR primers and the restriction enzymes NdeI and XhoI. The forward primer used was 5’-

TACGTCCATATGACTTCAGCACCTGCCATC-3’, and the reverse primer for CREB1-

Q2 was 5’-TAGCTGCTCGAGTCATGCTGCTTCTTCAGCAGG-3’ and for CREB1-Q2- bZIP was 5’-TAGCTGGCGGCCGCTCAATCTGATTTGTGGCAGTA-3’. ABL2-FABD was cloned from a pET-15 vector (clone HsCD00331895, PSI:Biology-Materials

Repository) using restriction enzymes, NdeI and XhoI. The pET-43. 1 Ek/LIC-OGT, pET-28b-CREB1-Q2, pET-28b-CREB1-Q2-bZIP, and pET-28b-ABL2FABD were each transformed into 10-beta E. coli cloning cells. Each plasmid was purified from a 5 mL

24 culture and sequenced (GeneLab DNA sequencing service at Louisiana State

University).

2. 2. 3 E. coli cultures and protein co-expression

BL21-Gold(DE3) E. coli competent cells were transformed with pET-43. 1-OGT

and a substrate vector (either pET-28b-CREB1-Q2, pET-28b-CREB1-Q2-bZIP, or pET-

28b-ABL2-FABD), and selected on agar plates supplemented with ampicillin and kanamycin. The expression strains were each grown in LB media at 37 °C. At an optical density at 600 nm of 0. 6–0. 8, the culture was induced with 0. 5 mM IPTG and grown at

37 °C for 3 h, unless otherwise stated. The cells were harvested using BugBuster protein extraction reagent and French press and/or ultrasonication followed by centrifugation at 30,000 g for 30 min at 4 °C. For some experiments, OGT and the substrate proteins were purified using His-Select Ni-affinity gel media using gravity flow.

Cleared lysate was loaded onto 1 mL of Ni-affinity media equilibrated with 50 mM sodium phosphate, 300 mM NaCl, pH 8. 0 buffer. The media was washed with 20 mL of buffer with 10 mM imidazole, and the protein was eluted with 10 mL of buffer with 300 mM imidazole. Having 6XHis-tags on both OGT and the substrate proteins allows for simultaneous purification and evaluation of protein expression levels using SDS-PAGE.

OGT ( 170 kDa) and the substrate proteins (<25 kDa) can be separated using gel filtration∼ chromatography if needed, or alternatively, OGT can be cloned without the

6XHis tag. Separation of OGT and the substrate proteins was not necessary for the experiments presented in this chapter.

25 BL21-Gold(DE3) E. coli with empty pET28b vector, K-12 parent E. coli, and K-12

∆nagZ E. coli were each grown in 250 mL of LB media at 37 °C. At an optical density at

600 nm of 0. 6, 1. 0 mM IPTG was added to the BL21-Gold(DE3) culture only and all

cultures were incubated for 3 h at 37 °C. The cells from each culture were harvested at

3000 g for 10 min, suspended in 12 mL of Bugbuster protein extraction reagent, and

lysed using French press. Cleared lysate was prepared by centrifugation at 30,000 g for

30 min.

2. 2. 4 O-GlcNAc detection and quantification

SDS-PAGE gels and blots were used to determine the relative amount of O-

GlcNAc-modified protein compared to total protein. Samples of cleared lysate and

insoluble pellet were prepared for SDS-PAGE in Laemmli buffer with 5% v/v 2- mercaptoethanol and incubated at 95 °C for 15 min. The samples were loaded into a 4–

20% gel and run at 125 V for 60 min in 25 mM tris, 192 mM glycine, 0. 1% SDS, pH 8. 3

buffer. Protein ladder was run on each gel, and duplicate gels were run when blot

analysis was needed. The gel was stained with Coomassie Brilliant Blue R250 solution

(0. 1% R250, 50% methanol, 40% water, 10% acetic acid), de-stained (50% water, 40%

methanol, 10% acetic acid), and imaged using a GE Typhoon 9410 imager or Fotodyne

Apprentice system. The contrast and brightness of the images were adjusted using

Microsoft Powerpoint software or ImageJ software to improve the quality of the images

for publication [48].

Blotting methods included both lectin and western blots. Fluorescein-conjugated

wheat-germ agglutinin was used for the lectin blots. Anti-O-GlcNAc MAb CTD 110. 6

and DyLight 550 conjugate Goat anti-Mouse IgG secondary antibody were used for the

26 western blots. Hen egg white lysozyme and ovalbumin were used as positive controls

for the western blots and lectin blots, respectively. All blots were imaged using the GE

Typhoon 9410 imager in the Louisiana State University AgCenter Biotechnology

Laboratory, Protein Facility. ImageJ software was used to measure the band densities

from the Coomassie-stained SDS-PAGE gels (relative total protein concentration) and

from the blots (relative O-GlcNAc-modified protein concentration) [48]. The background

signal was subtracted and the ratios of the blot to gel densities were calculated. The

error calculated from multiple background measurements was propagated and is

represented as error bars.

2. 2. 5 Partial purification of E. coli β-N-acetylglucosaminidase

E. coli β-N-acetylglucosaminidase was partially purified from the cleared lysate of

BL21-Gold(DE3) expressing empty pET-28b, K-12 parent, and K-12 ∆nagZ cultures

using cation exchange chromatography (CXC). A BioLogic DuoFlow 10 system with UV

detection and a fraction collector was used for all chromatography experiments. Cleared

lysate (12 mL) was separated using a UNO S1 cation exchange column into 1 mL

fractions using a buffer gradient of 20 mM sodium phosphate pH 7. 0, from 0 to 0. 5 M

NaCl. The fractions were analyzed for β-N-acetylglucosaminidase activity using the MU-

GlcNAc fluorescent well-plate assay described in Section 2. 6. The active fractions from

the cleared lysate of BL21-Gold(DE3) expressing empty pET-28b were combined and

used to determine enzyme kinetics and activity toward a glycopeptide. The K-12 parent

and K-12 ∆nagZ fractions were used to test for other β-N-acetylglucosaminidase activity toward MU-GlcNAc. The cleared lysate from the K-12 ∆nagZ culture was used to test

for other β-N-acetylglucosaminidase activity toward a glycopeptide.

27 2. 2. 6 E. coli β-N-acetylglucosaminidase detection assays

Spectrophotometric and fluorescence well-plate methods were used to measure

β-N-acetylglucosaminidase activity. A solution of a chromophoric, β-N- acetylglucosaminidase substrate, pNP-GlcNAc, was prepared in 50 mM tris-maleate,

20% glycerol, pH 7. 7 at 1. 0 mg/mL (2. 9 mM) and preheated to 37 °C. In a 96 wellplate, 20 µL of sample (cleared lysate or chromatography fraction) was added to 90

µL of the pNP-GlcNAc solution (final concentration was 2. 4 mM) and incubated at 37

°C for 30 min. The reaction was quenched with 200 µL of 42 mg/mL (396 mM) sodium carbonate, which stopped the reaction and converted any cleaved p-nitrophenol to yellow nitrophenolate. The absorbance at 405 nm was measured with a Biotek

PowerWave XS microplate spectrophotometer and converted to nmoles of nitrophenolate formed per mL of fraction/min using an extinction coefficient of 18,100 M-

1cm-1 for nitrophenolate [49]. An assay blank was run with 20 µL of buffer (lysis buffer or

chromatography buffer) in place of the sample; and, a sample blank was run with 90 µL of tris-maleate buffer in place of substrate. Each sample and blank were run in triplicate where indicated. For triplicate measurements, the average absorbance and standard deviation for each sample and blank were calculated. Detection of β-N- acetylglucosaminidase activity in the chromatography fractions were done by Maggie

Thomasson.

For the fluorescent β-N-acetylglucosaminidase activity assay, a stock solution of

4-methylumbelliferyl-β-N-acetylglucosaminide, MU-GlcNAc, was prepared at 25 mM in dimethylformamide. A working solution of 1. 5 mM MU-GlcNAc for well plate assays was made by diluting the stock solution in 100 mM sodium phosphate buffer, pH 7. 4.

28 The working solution was preheated to 37 °C. For the 96 well-plate assay, 20 µL of

sample (cleared lysate or chromatography fraction) was added to 90 µL of the MU-

GlcNAc working solution (final MU-GlcNAc concentration was 1. 2 mM) and incubated

at 37 °C for 30 min. The fluorescence intensity was measured with an excitation

wavelength of 372 nm and emission wavelength of 445nm using a Perkin Elmer Wallac

Victor2 V fluorescence plate reader, housed in the Louisiana State University AgCenter

Biotechnology Laboratory, Protein Facility. A positive control was run using 1 µL of β-N- acetylglucosaminidase from C. ensiformis (Jack bean) in 19 µL of buffer (lysis buffer or chromatography buffer) with 90 µL of the working MU-GlcNAc solution. A negative control was run using 20 µL of buffer (lysis buffer or chromatography buffer) with 90 µL of the working MU-GlcNAc solution. Each sample and blank were run in triplicate where indicated. For triplicate measurements, the average fluorescence intensity and standard deviation for each sample and blank were calculated. This fluorescent assay was done by Maggie Thomasson.

2. 2. 7 Kinetic assays

The Michaelis–Menten constants (Km) for MU-GlcNAc and pNP-GlcNAc and the inhibition constant (Ki) for PUGNAc were determined for the E. coli β-N- acetylglucosaminidase using the pooled β-N-acetylglucosaminidase-active CXC fractions (see Section 2. 5). For the pNP-GlcNAc reactions, three replicate samples of

0, 0. 05, 0. 5, 5 mM pNP-GlcNAc and 80 µL of active CXC fraction were mixed in 50 mM sodium phosphate, 500 mM sodium chloride, pH 7. 1 to a final volume of 100 µL. In addition to the sample blank (0 mM pNP-GlcNAc), an assay blank was made in triplicate with 5 mM pNP-GlcNAc in buffer. The samples were incubated at 37 °C. One set of the

29 triplicate samples was quenched with 200 µL of 396 mM sodium carbonate at 10, 15,

and 20 min. The formation of product was measured as described in Section 2. 6.

The Km for MU-GlcNAc and the Ki for PUGNAc were determined using sets of three replicate samples at 0, 13. 6, 16. 4, 20. 4, 27. 3, 40. 9, and 81. 8 µM MU-GlcNAc

with 0, 200, and 500 nM PUGNAc. Each sample volume was 110 µL with 20 µL of

active CXC fraction and 100 mM sodium phosphate, pH 7. 4 buffer. The fluorescence

intensities of the samples were measured at room temperature after 4, 9, and 14 min

using a fluorescence plate reader as described in Section 2. 6. The kinetics assay was

done by Maggie Thomasson.

2. 2. 8 Kinetic analysis

Initial velocities (V) at each substrate (S) concentration were determined using

least squares fitting to a linear function, and the average values were calculated for

triplicate measurements. The data were then analyzed by nonlinear regression analysis

using the computer programs of Cleland (1979). Initial velocities were fit to Eq. (1) to

determine the maximal velocity (Vmax) and the Michaelis–Menten constant (Km).

Eq. (1)

Data for competitive inhibition were fit to Eq. (2), where I is the concentration of

PUGNAc and Ki is the inhibition constant. The kinetics analysis was done by Maggie

Thomasson.

30 Eq. (2)

2. 2. 9 Inhibition of E. coli β-N-acetylglucosaminidase activity

Cleared lysate from a BL21-Gold(DE3) culture was tested for β-N- acetylglucosaminidase activity using the pNP-GlcNAc spectrophotometric assay described in Section 2. 6. PUGNAc was added to 20 µL of lysate at varying concentrations (0, 1, and 2 µM PUGNAc) and mixed with 90 µL of 1. 0 mg/mL pNP-

GlcNAc. Conditions, blanks, and measurements are the same as described in Section

2. 6.

Changes in the relative amount of O-GlcNAc-ABL2-FABD were measured for co- expression cultures grown with 0, 1, 2, and 10 µM PUGNAc. PUGNAc was added to the growth media of BL21-Gold(DE3) cells co-expressing pET43. 1 Ek/LIC-OGT and pET28b-ABL2-FABD. The cultures were grown, harvested, and lysed as described in

Section 2. 3. The soluble and insoluble proteins from each culture were analyzed with

SDS-PAGE and western blot as described in Section 2. 4. The relative amount of O-

GlcNAc was measured as described in Section 2. 4.

2. 2. 10 NMR experiments to detect E. coli β-N-acetylglucosaminidase hydrolysis of an O-GlcNAc-modified peptide

To evaluate the activity of the E. coli β-N-acetylglucosaminidase toward protein

O-GlcNAc, a synthetic O-GlcNAc-modified peptide (SVES(β-O-GlcNAc)GSADAK-NH2) was incubated with partially purified, E. coli β-N-acetylglucosaminidase and K-12 ∆nagZ lysate. Each reaction mixture was analyzed with 1H NMR to observe the 1H anomeric

31 signal of any released α-GlcNAc (5. 2 ppm). This signal was chosen because it has

relatively little background interference. The reaction mixtures contained 200 µM

glycopeptide, 100 µL of β-N-acetylglucosaminidase-active CXC fraction or K-12 ∆nagZ

lysate, 50 mM sodium phosphate, 500 mM NaCl, pH 7. 1 buffer, 10% D2O, and trace

DSS (to lock and reference the NMR signal, respectively).

A glycopeptide control, sample control, and sample control with 200 µM N- acetylglucosamine (GlcNAc) were prepared with 100 µL of CXC buffer instead of active fraction or lysate, 3. 0 µL of water instead of glycopeptide stock solution, and 1. 3 µL of

20 mM GlcNAc plus 2. 7 µL of water instead of glycopeptide stock solution, respectively.

The samples were either incubated at 37 °C for 5 h (CXC β-N-acetylglucosaminidase- active fraction) or 22. 5 h (K-12 ∆nagZ lysate). The glycopeptide control and sample controls were treated with the same conditions. The reactions were performed one time.

The reaction mixtures and controls were analyzed in 4 mm D2O-susceptibility matched

tubes. 1H NMR spectra of the samples were collected on a 700 MHz Varian NMR

equipped with a 5 mm HCN probe. The spectra were collected at 20 °C with

presaturation of water (1 s saturation time, 12 dB saturation power), 1 s acquisition time,

1 s recycle delay, 7000 Hz spectral width, 30 dB gain, and 2048 scans. Data

processing, line broadening (1 Hz exponential), chemical shift referencing, baseline

correction, peak height measurements, and signal-to-noise calculations were performed

with VnmrJ 3. 2 software (Agilent). The amount of released GlcNAc was estimated by

comparing the peak intensities at 5. 198 ppm (left peak of the anomeric 1H α-GlcNAc doublet) of the reaction mixture to the control sample with 200 µM GlcNAc. For the glycopeptide reaction with K-12 ∆nagZ lysate, the limit of quantitation (S/N = 10) was

32 calculated using the S/N of the 1H anomeric α-GlcNAc peak in the control sample with

200 µM GlcNAc.

2. 3 Results and discussion

2. 3. 1 Co-expression of OGT with substrate proteins using E. coli

The second glutamine-rich domain of human CREB1 (CREB1-Q2) and the Q2 and bZIP domains of CREB1 (CREB1-Q2-bZIP) were co-expressed with full-length human OGT in BL21-Gold(DE3) E. coli. Even though these vectors have the same origin of replication, the different antibiotic resistance prevents incompatibility [50]. The

CREB1-Q2 and CREB1-Q2-bZIP domains were chosen as substrate proteins because the Q2 domain is natively modified with at least two O-GlcNAc molecules that suppress

CREB1-activated transcription [17]. The co-expression cultures were lysed and

separated into soluble lysate and insoluble pellet fractions. The soluble lysate and pellet

were analyzed with Coomassie SDS-PAGE and lectin blot. The results for the co-

expression of OGT with CREB1-Q2 and CREB1-Q2-bZIP are shown in Figures 2-2 and

2-3, respectively. Figures 2-2a and 2-3a show the SDS-PAGE gel images for Ni-affinity

purified CREB1-Q2 and CREB1-Q2-bZIP soluble lysate. CREB1-Q2 is highly expressed and soluble, while CREB1-Q2-bZIP is mainly found in the pellet. The additional bands in

the Ni-affinity purified lysate in Figure 2-2a are E. coli background proteins and 6XHis-

tagged OGT (with multiple translational truncations due to rare codons). Figures 2-2b

and 2-3b show the lectin blots corresponding to the SDS-PAGE gels. The dark bands

indicate where FITC-WGA bound O-GlcNAc-modified proteins. The presence of O-

GlcNAc was also confirmed with an anti-O-GlcNAc MAb CTD 110. 6 western blot of Ni-

affinity purified protein from the expression of CREB1-Q2-bZIP (pET-28b-CREB1-Q2-

33 bZIP) without and with co-expression of OGT (pET-43. 1-OGT) in E. coli BL21-

Gold(DE3) (see Supplementary Figure 1). In both lectin blots (Figures 2-2 and 2-3), only insoluble CREB1-Q2 and CREB1-Q2-bZIP were modified with O-GlcNAc.

The absence of soluble, O-GlcNAc-modified protein from the co-expression system was unexpected. Others have shown that E. coli co-expression of OGT with substrate protein produce O-GlcNAc-modified proteins in vivo, but the solubility of the modified proteins was not specifically addressed in these studies [51-53].

Figure 2-2 The Ni-affinity purified soluble lysate and pellet from co-expression of CREB1-Q2 (pET-28b-CREB1-Q2) and OGT (pET-43. 1-OGT) in E. coli BL21- Gold(DE3) analyzed with (a) Coomassie-stained SDS-PAGE and (b) FITC-WGA lectin blot detecting O-GlcNAc. The bands corresponding to CREB1-Q2 are indicated with asterisks.

Figure 2-3 The soluble lysate and pellet from co-expression of CREB1-Q2-bZIP (pET28b-CREB1-Q2-bZIP) and OGT (pET-43. 1-OGT) in E. coli BL21-Gold(DE3) analyzed with (a) Coomassie-stained SDS-PAGE and (b) FITC-WGA lectin blot detecting O-GlcNAc. The bands corresponding to CREB1-Q2-bZIP are indicated with asterisks. From Figure 2-3a, CREB1-Q2-bZIP shows little soluble protein in the SDS-PAGE image, so observing only O-GlcNAc-modified CREB1-Q2-bZIP in the pellet is not informative. However, the solubility of the CREB1-Q2 domain is good, but the O-

GlcNAc-modified protein only occurs in the pellet. To improve the solubility of the O-

GlcNAc-modified proteins, the co-expression of OGT with CREB1-Q2 and CREB1-Q2- bZIP was tested using additional E. coli expression strains at various conditions.

Rosetta, C41, and C43 E. coli strains were induced with 0. 5 mM IPTG and grown at 37

°C for 3 h and 18 °C for 16 h. The Tuner E. coli strain was grown at the same

temperatures but with 0. 25 mM–1 mM IPTG. Arctic express E. coli cells were induced

with 0. 5 mM IPTG and grown at 4 °C and 12 °C for 24 h. Regardless of the type of E.

coli expression strain or conditions used, the O-GlcNAc-modified CREB1-Q2-bZIP and

CREB1-Q2 proteins were only found in the pellet (data not shown).

35 To circumvent the possibility that O-GlcNAc-modified CREB1-Q2 and CREB1-

Q2-bZIP are inherently insoluble, the F-actin binding domain of Abelson tyrosine-kinase

2 (ABL2-FABD, also known as ARG, Abelson-related gene protein) was selected as a third test protein. ABL2-FABD was selected because it expresses in high yield with high solubility in E. coli, actin-binding proteins have been shown to be regulated by O-

GlcNAc, and ABL2-FABD is predicted to have two sites of O-GlcNAc modification [54-

57].

Figure 2-4 ABL2-FABD (pET-28b-ABL2-FABD) and OGT (pET-43. 1-OGT) were co- expressed in E. coli BL21-Gold(DE3) with increasing concentrations of PUGNAc in the

36 growth media (0, 1, 2, and 10 µM PUGNAc). The soluble lysate (S) and pellet (P) from each culture was analyzed with (a) Coomassie-stained SDS-PAGE and (b) anti-O- GlcNAc MAb CTD 110. 6 western blot. HEWL was used as a positive control in the last lane. (c) A plot of the relative band density of the western blot band to the Coomassie- stained SDS-PAGE gel band (indicated with an asterisk). For each band in (a) and (b), the background density was measured thrice, subtracted from each band, and used to estimate the propagated measurement error (represented as error bars in (c)). Figure 2-4a (Lanes 2 and 3) shows an SDS-PAGE image of the soluble lysate

and pellet from ABL2-FABD co-expressed with OGT. The ABL2-FABD protein is

abundant in both the lysate and pellet, but the western blot image in Figure 2-4b (Lanes

2 and 3) shows small amounts of O-GlcNAc-modified ABL2-FABD only in the pellet.

The absence of soluble O-GlcNAc-modified protein from the co-expression system

suggests two possibilities: (1) the modification causes the proteins to precipitate and/or

(2) the O-GlcNAc modification is being cleaved by an endogenous E. coli enzyme in the cytoplasm. The second possibility assumes that the insoluble, O-GlcNAc-modified proteins are protected from the suspected E. coli β-N-acetylglucosaminidase activity.

2. 3. 2 Inhibiting NagZ activity in E. coli increases in vivo O-GlcNAc modification

NagZ, the only E. coli β-N-acetylglucosaminidase identified, is suspected to be

the enzyme that deglycosylates soluble O-GlcNAc-modified protein from the co-

expression system. The existence of an O-GlcNAc transferase in E. coli has been ruled

out but the E. coli β-N-acetylglucosaminidase, NagZ, has not been tested for activity

toward O-GlcNAc-modified peptides or proteins [58]. NagZ was first purified from K-12

E. coli and characterized using pNP-GlcNAc in a spectrophotometric assay [59]. The

soluble, cytoplasmic enzyme was found to specifically cleave O-β-linked GlcNAc. Later

characterization of K-12 NagZ demonstrated its role in peptidoglycan recycling by

cleaving the GlcNAc-(1,4)-MurNAc bond (MurNAc, N-acetylmuramic acid) with a Km of

37 35 µM for the anhydro-disaccharide [60, 61]. A null mutant showed that NagZ was not

essential to E. coli growth and morphology and was the only E. coli β-N-

acetylglucosaminidase [59, 60]. It is possible that NagZ has activity toward O-GlcNAc-

modified proteins because it is similar to human O-GlcNAcase (OGA) in substrate (pNP-

GlcNAc) and inhibitor (PUGNAc). Human OGA cleaves pNP-GlcNAc and is inhibited by

PUGNAc [62, 63]. K-12 NagZ, which has 99% amino acid identity to BL21-Gold(DE3)

NagZ, has a pNP-GlcNAc Km of 0. 31–0. 43 mM [59, 61]. PUGNAc is a potent inhibitor

of Vibrio cholerae NagZ (Ki = 48 nM), a homolog of E. coli NagZ with 56% amino acid

sequence identity [64]. Following the hypothesis that NagZ cleaves O-GlcNAc from

modified proteins in the co-expression system, PUGNAc was tested as an inhibitor of

the enzyme in vitro and in vivo.

The MU-GlcNAc Km and the PUGNAc Ki of BL21-Gold(DE3) E. coli β-N-

acetylglucosaminidase were determined using partially purified β-N-

acetylglucosaminidase from BL21-Gold(DE3) lysate. β-N-Acetylglucosaminidase-active

CXC fractions were identified with the MU-GlcNAc well-plate assay and pooled together

(Supplementary Figure 2). The MU-GlcNAc Km and the PUGNAc Ki were determined by

measuring the formation rate of 4-methylumbelliferone with a fluorescence plate reader at varying MU-GlcNAc and PUGNAc concentrations. The data shown in Figure 2-5 were

globally fit to competitive, noncompetitive, and uncompetitive inhibition models using

nonlinear regression analysis, and the competitive model gave the best fit. This result is

expected since PUGNAc is a competitive inhibitor of other β-N-acetylglucosaminidases and has been shown to bind at V. cholerae NagZ’s active site [62, 64, 65]. The best fit

to Eqs. (1) and (2) give a Km of 5. 6 ± 1. 2 µM MU-GlcNAc and a Ki of 25. 5 ± 4. 6 nM

38 PUGNAc. These results are similar to the Km of K-12 NagZ (35 µM anhydro-

disaccharide) and the Ki of V. cholerae NagZ (Ki = 48 nM PUGNAc), respectively [61,

64].

Even though PUGNAc is a strong inhibitor of partially purified E. coli β-N-

acetylglucosaminidase, the effectiveness of the inhibitor is reduced in cleared lysate.

The E. coli β-N-acetylglucosaminidase activity in cleared lysate was tested with the

addition of 0, 1, and 2 µM PUGNAc.

Figure 2-5 The initial velocities of E. coli β-N-acetylglucosaminidase, partially purified with CXC from BL21-Gold(DE3) cleared lysate, at varying MU-GlcNAc concentrations and with different concentrations of PUGNAc. The points represent the observed velocities, and the lines represent the best fits of the data to Eqs. (1) and (2). The results of this assay (Figure 2-6) show that PUGNAc inhibits the β-N-

acetylglucosaminidase activity, but not as well as expected for a Ki of 25. 5 nM. To rule

out the possibility that E. coli β-N-acetylglucosaminidase has a low pNP-GlcNAc Km, the

Km was measured using partially purified β-N-acetylglucosaminidase from BL21-

Gold(DE3) lysate. The best fit of the data (Supplementary Figure 3) to Eq. (1) results in

a Km of 0. 35 ± 0. 05 mM, which is within the range of reported values for K-12 NagZ (0.

31–0. 43 mM) and does not explain the reduced effectiveness of PUGNAc in the

39 cleared lysate [59, 61]. Another E. coli β-N-acetylglucosaminidase with activity for pNP-

GlcNAc could also explain these results, but this hypothesis is ruled-out by the lack of activity of the null mutant (Section 3. 4) and by work published previously [60]. Another

possibility is that the effective concentration of PUGNAc is reduced in the lysate

compared to the CXC active fraction; other enzymes or cellular material in the lysate

may sequester or degrade PUGNAc.

To assess whether inhibition of the E. coli β-N-acetylglucosaminidase increases

the amount of O-GlcNAc-modified ABL2-FABD produced with the co-expression

system, PUGNAc was added to the growth media at 0, 1, 2, and 10 µM. The relative

amount of O-GlcNAc-modified ABL2-FABD produced was determined by normalizing

the western blot band density to total ABL2-FABD protein using the band density of a

duplicate SDS-PAGE gel image (Figure 2-4a and b). As shown Figure 2-4c, the amount

of insoluble, O-GlcNAc-modified ABL2-FABD increases with increasing PUGNAc,

consistent with an E. coli β-N-acetylglucosaminidase cleaving O-GlcNAc in vivo. Even

with 10 µM PUGNAc, soluble O-GlcNAc-modified ABL2-FABD is absent (Figure 2-4b).

This result is consistent with poor inhibition of E. coli β-N-acetylglucosaminidase in the

cleared lysate with PUGNAc. The possible explanations are the same as discussed in

the previous paragraph with the additional possibility of inefficient transport into the cell.

Alternatively, the results suggest that soluble, O-GlcNAc-modified ABL2-FABD is

unstable. However, the latter is not supported by the fact that the relative levels of

soluble to insoluble ABL2-FABD (as assessed by SDS-PAGE (Figure 2-4a)) are not

significantly different even when O-GlcNAc is increased. These results taken together

indicate that an E. coli enzyme, likely NagZ, cleaves O-GlcNAc from the soluble protein,

40 the enzyme is inhibited by PUGNAc, the effectiveness of PUGNAc inhibition in cleared

lysate and in vivo is reduced, and the insoluble O-GlcNAc-modified ABL2-FABD is protected from the β-N-acetylglucosaminidase activity.

Figure 2-6 The β-N-acetylglucosaminidase activity of cleared lysate with PUGNAc (0, 1, 2 µM) was measured using the pNP-GlcNAc spectrophotometric assay. The lysate was from the co-expression of ABL2-FABD (pET-28b-ABL2-FABD) and OGT (pET43. 1- OGT) in E. coli BL21-Gold(DE3). Each sample and control were run in triplicate, and the error bars represent the standard deviation of these measurements. The 405 nm absorbance values were converted to activity (right vertical axis) using the extinction coefficient of nitrophenolate.

2. 3. 3 E. coli NagZ cleaves O-GlcNAc from a peptide

To determine if E. coli β-N-acetylglucosaminidase can deglycosylate an O-

GlcNAc-modified peptide, the release of GlcNAc from a synthetic O-GlcNAc-modified

peptide incubated with partially purified E. coli β-N-acetylglucosaminidase was

monitored with 1H NMR. The α-GlcNAc anomeric 1H signal at 5. 2ppm (Figure 2-7, full

spectra in Supplementary Figure 4) was monitored for the glycopeptide control, the

glycopeptide and active fraction mixture, the active fraction control, and the active

fraction control with 200 µM GlcNAc after incubation at 37 °C for 5 h. No α-GlcNAc was

detected in the glycopeptide or active fraction control samples (Figure 2-7a and d). α-

GlcNAc was detected in the mixture of glycopeptide and partially purified β-N-

acetylglucosaminidase sample (Figure 2-7c) at a concentration of approximately 40 µM

41 GlcNAc. These results demonstrate that the E. coli β-N-acetylglucosaminidase has

activity toward this O-GlcNAc-modified peptide, and likely has activity toward O-GlcNAc-

modified proteins in vivo.

Figure 2-7 Fractions from CXC of BL21-Gold(DE3) lysate with β-N- acetylglucosaminidase activity toward MU-GlcNAc were pooled and tested for β-N- acetylglucosaminidase activity toward the synthetic O-GlcNAc-modified peptide, SVES(β-O-GlcNAc)GSADAK-NH2. (a) CXC fraction control, (b) CXC fraction control with 200 µM GlcNAc added, (c) CXC fraction with 200 µM glycopeptide, and (d) 200 µM glycopeptide control in CXC buffer. The spectra shown are expanded regions from 5. 1 to 5. 3 ppm, with a gray dashed line indicating the α-GlcNAc chemical shift (center of the doublet). All spectra were processed and expanded identically for direct comparison. The full spectra are included in Supplementary Figure 4. The glycopeptide control sample (d) has extra noise due to truncation artifacts from the water peak; the additional noise is small and does not affect spectral interpretation. Based on the 1H

42 anomeric α-GlcNAc intensity in (b), approximately 40 µM GlcNAc was hydrolyzed by the E. coli β-N-acetylglucosaminidase in 5 h at 37 °C (c).

2. 3. 4 Identifying the E. coli β-N-acetylglucosaminidase as NagZ

To verify that NagZ is the only β-N-acetylglucosaminidase in E. coli and the enzyme purified and characterized in these studies, a K-12 ∆nagZ strain and parent strain were tested for β-N-acetylglucosaminidase activity. Cleared lysate from cultures were tested for activity using the MU-GlcNAc fluorescence assay (Supplementary

Figure 5). As expected, activity was detected in the K-12 parent strain. No significant activity was found in any of the fractions or the cleared lysate of the K-12 ∆nagZ strain, indicating that NagZ is the only β-N-acetylglucosaminidase in E. coli, as Yem and Wu

also observed [59]. These results confirm that the β-N-acetylglucosaminidase-active

fractions used in the kinetic assays and NMR experiments must have contained NagZ.

To determine if NagZ is the only E. coli enzyme with activity toward O-GlcNAc-

modified proteins, lysate from K-12 ∆nagZ strain was incubated with 200 µM O-GlcNAc-

modified peptide and monitored with 1H NMR. Even after 22. 5 h of incubation at 37 °C,

the 1H spectrum (Figure 2-8c, full spectra in Supplementary Figure 6) did not show any

detectable α-GlcNAc greater than the limit of quantitation (10 µM). The lack of activity

toward the glycopeptide indicates that NagZ is the one and only enzyme that sabotages

the in vivo production of O-linked β-N-acetylglucosamine-modified proteins with E. coli.

Figure 2-8 1H NMR spectra of (a) ∆nagZ K12 lysate, (b) lysate with 200 µM GlcNAc, and (c) lysate with 200 µM glycopeptide, SVES(β-O-GlcNAc)GSADAK-NH2. All samples were incubated at 37 °C for 22. 5 h to test for β-N-acetylglucosaminidase activity toward the synthetic O-GlcNAc-modified peptide. The spectra shown are expanded regions from 5. 1 to 5. 3 ppm, with a gray dashed line indicating the α-GlcNAc chemical shift (center of the doublet). All spectra were processed and expanded identically for direct comparison. The full spectra are included in Supplementary Figure 6. Based on the signal-to-noise ratio of the 200 µM GlcNAc peak in (b), the limit of quantitation (LOQ) in (c) is 10 µM GlcNAc.

2. 4 Conclusions

E. coli has not been used to produce milligram quantities of soluble O-GlcNAc- modified proteins because an endogenous E. coli enzyme, NagZ, removes O-GlcNAc from modified proteins in vivo and in vitro. An E. coli co-expression system was developed to modify proteins with O-GlcNAc in vivo using human OGT. E. coli is a good

44 host because it lacks endogenous OGT activity, can produce milligram quantities of

protein, and allows for metabolic labeling with isotopes for NMR structure determination.

Three substrate proteins were successfully O-GlcNAc-modified with the co-expression

system: CREB-Q2, CREB-Q2-bZIP, and ABL2-FABD. CREB-Q2 and ABL2-FABD, in

particular, are soluble proteins, but the O-GlcNAc-modified proteins were only found in

the insoluble pellet. The O-GlcNAc-modification may cause the substrate proteins to

become unstable and precipitate; particularly, if the modification is added co-

translationally before the protein has folded. Independent of whether O-GlcNAc causes insolubility, several pieces of evidence suggest that E. coli NagZ cleaves O-GlcNAc

from the soluble substrate protein in vivo and in vitro. The amount of O-GlcNAc-

modified ABL2-FABD produced with the co-expression system increased when a β-N-

acetylglucosaminidase inhibitor, PUGNAc, was added to the growth media. NMR was

used to show that an E. coli enzyme cleaves O-GlcNAc from an O-GlcNAc-modified

peptide, a foreign substrate. Even though the E. coli β-N-acetylglucosaminidase was not

directly identified with mass spectrometry, the lack of β-N-acetylglucosaminidase activity

in lysate from a K-12 ∆nagZ strain both toward MU-GlcNAc and an O-GlcNAc-modified peptide demonstrates that NagZ is the only E. coli β-N-acetylglucosaminidase.

Moreover, the lack of activity in the knockout confirms that NagZ was present in the samples shown to cleave O-GlcNAc and is inhibited by PUGNAc. The Km’s of

BL21Gold(DE3) NagZ determined, 0. 35 mM pNP-GlcNAc and 5. 6 µM MU-GlcNAc, are

consistent with values published for K-12 NagZ [59, 61]. The inhibition constant of

BL21-Gold(DE3) NagZ was determined to be 25. 5 nM and is consistent with V.

cholerae NagZ, but the mechanism of reduced PUGNAc effectiveness in cleared lysate

45 and in in vivo experiments remains unclear [64]. Efforts are currently underway to test

the co-expression system in ∆nagZ E. coli strains for increased yields and soluble O-

GlcNAc-modified proteins.

2. 5 Acknowledgments

We thank Dr. John Hanover for his gift of the pET-43. 1 Ek/LIC-OGT vector. A. L.

thanks the Louisiana State University HHMI Summer Undergraduate Research Program

for support. M. M. thanks Prof. Aaron Smith for proof-reading the manuscript, Prof.

Grover Waldrop for advice on enzyme kinetics, Dr. Tyler Broussard for assistance with the global kinetic fitting software, and the Louisiana State University Structural Biology

Group for helpful suggestions regarding experiments. This research was funded by

Louisiana State University.

46 Chapter 3: Is NagZ the saboteur?

Chapter 3. 1 Introduction

3. 1. 1 The Hexosamine biosynthesis pathway in E. coli

UDP-GlcNAc is an important component in E. coli. It is a major precursor of the

bacterial cell wall. (3) The enzymes of the hexosamine biosynthesis pathway in E. coli

are different from those in humans, but the steps of the pathway are similar. Glucose is

transported into the E. coli cell and phosphorylated by a glucose specific

phosphotransferase (PstG), and phospho-glucose isomerase (Pgi) interconverts

glucose-6-phosphate to fructose-6-phospate. Glucosamine-6-phosphate synthase

(GlmS) transfers an amino group from glutamine to make glucosamine-6-phosphate,

which is then converted to glucosamine-1-phosphate by phosphoglucosamine mutase

(GlmM). N-acetylglucosamine-1-phosphate is formed through the acetylation of

glucosamine-1-phosphate by glucosamine-1-phosphate acetyltransferase (GlmU).

Finally, N-acetylglucosamine-1-phosphate is uridylated by N-acetylglucosamine-1- phosphate uridylyltransferase (GlmU) to form uridine diphosphate N-acetylglucosamine

(UDP-GlcNAc). Glucosamine and N-acetylglucosamine can enter the HBP as well.

Glucosamine is transported into the cell and phosphorylated by mannose transporter

(ManXYZ), and N-acetylglucosamine is transported and phosphorylated by NagE, a N- acetylglucosamine specific transporter. N-acetylglucosamine is deacetylated by N- aceytglucosamine-6-phosphate deacetylase to glucosamine-6-phosphate where it continues along the HBP [66, 67]. N-acetylglucosamine that is cleaved from the cell wall by NagZ is phosphorylated by NagK and used to make more UDP-GlcNAc [68].

Figure 3-1 Glucose, glucosamine, and N-acetylglucosamine are used by E. coli’s hexosamine biosynthesis pathway to make UDP-GlcNAc.

3. 1. 2 The role of β-N-acetylglucosaminidase (NagZ) in E. coli

E. coli is a prokaryotic single cell organism with a cellular membrane made of

murein also known as peptidoglycan. Peptidoglycan is composed of cross-linked alternating GlcNAc-β1,4-MurNAc-peptide monomers [68]. The main function of the

peptidoglycan layer is to support the cell wall structure and balance the cytoplasm’s

48 osmotic pressure. There are several enzymes that participate in the degradation of the

peptidoglycan layer. NagZ is the only β-N-acetylglucosaminidase whose function is to

cleave N-acetylglucosamine (GlcNAc) from N-acetylmuramic acid (MurNAc) peptides for peptidoglycan recycling [69].

Figure 3-2 NagZ cleaves the bond between GlcNAc and MurNAc, of the peptidoglycan layer [70]. Adapted from: Triassi, A. J. , M. S. Wheatley, M. Savka, H. M. Gan, R. C. J. Dobson, and A. O. B. Hudson, L,l-diaminopimelate aminotransferase (dapl): A putative target for the development of narrow-spectrum antibacterial compounds. Frontiers in Microbiology, 2014. 5. GlcNAc, clea:ved from the cell, is recycled to make more UDP-GlcNAc, which is

a precursor used to make the peptidoglycan. Knocking out nagZ inhibits the cells ability

49 to cleave GlcNAc from MurNAc leading to a build-up of the disaccharide in the cytoplasm. nagZ knockout (ΔnagZ) cells have less GlcNAc and UDP-GlcNAc in the cytoplasm compared to the wild type [69]. The level of O-GlcNAcylation is dependent upon the amount of UDP-GlcNAc [71]. Although a ∆nagZ system for expression of O-

GlcNAc-modified proteins may prevent glycosidase activity, it may be limited in the amount of glycoprotein it can produce because of the lower level of UDP-GlcNAc. In this chapter we will analyze the effects of ΔnagZ on the expression of O-GlcNAc-modified proteins.

3. 2 Materials and Methods

3. 2. 1 Materials

Hen egg white lysozyme, glycerol, glucose (Glc), glucosamine (GlcN), N-acetyl-

D-glucosamine (GlcNAc), O-(2-acetamido-2-dexoy-d-glucopyranosylidene)amino-N- phenylcarbamate (PUGNAc), sodium lauroyl sarcosinate (sarkosyl), sodium chloride

(NaCl), sodium phosphate dibasic, sodium phosphate monobasic, 4-(2-hydroxyethyl)-1- piperazineethanesulfonic acid (HEPES), bromophenol blue, Coomassie brilliant blue

R250, sodium dodecyl sulfate (SDS), glycine, tris(hydroxymethyl)aminomethane (tris),

2-mercaptoethanol, imidazole, urea, acetic acid, methanol, and kanamycin were purchased from Sigma-Aldrich. Rosetta(DE3)pLysS cells, BugBuster protein extraction reagent, and ampicillin were purchased from EMD Millipore/Novagen. A UNO S1 cation exchange column, Precision Plus Protein Unstained Standards, 4–20% mini-PROTEAN

TGX precast gels, immun-blot low fluorescence PVDF membrane, and Laemmli buffer were purchased from Bio-Rad. 2-log DNA ladder, 10-beta cells, T4 DNA ligase, NdeI restriction enzyme, and XhoI restriction enzyme were purchased from New England

50 Biolabs. BioReagents EZ-Run Rec Protein Ladder, Coomassie (Bradford) Protein Assay kit, O-GlcNAc-modified BSA, anti-O-GlcNAc MAb CTD 110. 6, and DyLight 550- conjugate Goat anti-Mouse IgG secondary antibody were purchased from Thermo

Scientific/Pierce Antibodies. Luria broth (LB) and agar were purchased from BD

Biosciences. Wheat germ agglutinin (WGA) agarose and O-GlcNAc-modified BSA were purchased from Vector Labs. Other chemicals and materials used include isopropyl β-

D-1-thiogalactopyranoside (IPTG, Affymetrix Anatrace), Ni Sepharose 6 Fast flow (GE

Healthcare), chloramphenicol (Calbiochem), QIAprep spin miniprep kit (Qiagen), Costar

96 well UV transparent flat bottom plates, K-12 ∆nagZ knockout strain (JW1093-1) and

K-12 parent strain (BW25113) (Keio Collection, E. coli Genetic Resources at Yale,

CGSC, The Coli Genetic Stock Center) [72], and clone HsCD00331895 of the F-actin binding domain of Abelson tyrosine-kinase 2 (ABL2-FABD) in pET15 vector (DNASU

Plasmid Repository, PSI:Biology-Materials Repository) [46]. All water used was supplied from a Millipore Direct-Q 3 ultrapure water system.

3. 2. 2 Knocking out the nagZ gene in the Rosetta(DE3) pLysS strain

The ∆nagZ Rosetta(DE3)pLysS strain was made by Dr. William Doerrler and Lisa

Boughner, a graduate student.

3. 2. 3 Making the ΔnagZ Rosetta(DE3) pLysS cells competent

Four single ∆nagZ Rosetta(DE3)pLysS colonies were cultured in 5 mL of LB with kanamycin (50 µg/ml) and chloramphenicol (35 µg/ml) at 37 oC overnight. These cultures were added to 50 mL of fresh LB with the appropriate antibiotics and cultured for 3 hr. The cells were cooled on ice for 20 min and spun down at 3,000 g for 15 min at

51 o o 4 C. The cells were resuspended in 10 mL of 0. 1 M CaCl2 at 4 C and spun down

again. The supernatant was removed and the cells were allowed to dry for 1 min. The

o cells were reconstituted in 2 mL of 0. 1 M CaCl2 at 4 C and stored as 100 µl aliquots for

transformation.

3. 2. 4 Transformation

pCDF-ABL2 and pET43. 1-OGT were transformed into the ΔnagZ

Rosetta(DE3)pLysS competent cells using a heat shock method. A mixture of 1 µl of

each plasmid and 100 µl of the competent cells was heat shocked at 42 oC for 45 s. The

cells were immediately chilled on ice for 2 min and incubated with shaking for 1 hr at 37

oC. ΔnagZ Rosetta(DE3)pLysS colonies with pCDF-ABL2 and pET43. 1-OGT were grown on LB agar with 50 µg/mL spectinomycin, 100 µg/mL ampicillin, 50 µg/mL kanamycin, and 35 µg/mL chloramphenicol.

3. 2. 5 Expression conditions

ABL2 and OGT were co-expressed in ΔnagZ Rosetta(DE3)pLysS cells. Three

colonies of ΔnagZ Rosetta(DE3)pLysS with pCDF-ABL2 and pET43. 1-OGT were

cultured overnight at 37 oC in 150 mL of LB with the appropriate antibiotics. 1 mL of the

overnight culture was used to inoculate fresh LB or minimal media (see Table 3-1 for

minimal media expression recipe) with the appropriate antibodies at 37 oC. Once the

cultures reached an OD600 of 0. 6-0. 8, they were divided into three 50 mL aliquots and

induced with 50 µL of 0. 5 M IPTG. Each 50 mL aliquot was incubated at 16 oC, 23 oC,

and 37 oC for 16 hr, 12 hr, and 3 hr, respectively. The cells were harvested by

centrifugation at 3000 g for 15 min at 4 oC. The supernatant was discarded, and the

52 pellet was reconstituted with 1 mL of BugBuster. The suspended cells were sonicated using a Sonics Vibra-cell VCX130 at 30% power for 120 s (3 cycles of 30 s on/10 s off).

The lysed cells were spun down at 30,000 g for 30 min at 4 oC. The cleared lysate was separated from the pellet and saved for nickel affinity chromatography. The expression procedure was performed for pCDF-ABL and pETDuet-1-codon optimized OGT co- expression in ΔnagZ Rosetta(DE3)pLysS.

Table 3-1 The recipe for 1 L of minimal growth media, 1 L of 10X M9 salts, and 100 mL of 1000X trace mineral.

10X M9 salts 100. 0 ml

D-glucose stock (20 g/100 ml) (0. 2 µm filter sterilized) 20. 0 ml

1000X Trace minerals 1. 0 ml

1 M MgCl2 (autoclaved) 2. 0 ml

50 mM CaCl2 (autoclaved) 2. 0 ml

Vitamin Solution (1 tablet/20 ml) (0. 2 µm filter sterilized) 2 ml

10X M9 salts (autoclaved)

Na2HPO4•7H2O 128 g

KH2PO4 30 g NaCl 5 g

NH4Cl 10g

1000X Trace Minerals (10 mM each) (autoclaved)

FeCl3•6H2O 0. 2703 g

CuSO4•5H2O 0. 2497 g

MnSO4•6H2O 0. 1690 g

ZnSO4•6H2O 0. 2875 g

53 3. 2. 6 Ni-affinity purification conditions

Ni Sepharose 6 Fast Flow with a binding capacity of 40 mg/mL was used in 0. 1 mL aliquots. The nickel media was placed in microspin columns and centrifuged at 300 g to remove the storage buffer. Water was used to rinse the media, followed by 1 mL of

wash buffer (50 mM phosphate, 150 mM NaCl, and 20 mM imidazole at pH 8) to

equilibrate the media. The cleared lysate was added to the equilibrated media and

allowed to flow through by gravity. The media was washed with 2 mL of wash buffer.

The protein of interest was eluted using 200 µL of elution buffer (50 mM phosphate, 150

mM NaCl, and 300 mM imidazole at pH 8).

3. 2. 7 SDS-PAGE and western blot analysis

The cleared lysate and insoluble pellet were analyzed with SDS-PAGE and

western blot analysis. The cleared lysate was diluted in a 1:5 ratio with BugBuster or

SDS-PAGE buffer. 20 μL of the dilute cleared lysate was added to 20 μL of Laemmli

sample buffer. The samples were divided in half to run on two different gels. Purified

samples were prepared as 30 μL aliquots in a 1:1 ratio with Laemmli buffer. These

samples were also divided for two different gels: one gel was stained with Coomassie

brilliant blue R-250 stain, and the other gel was prepared for O-GlcNAc detection by

western blot analysis. All samples were boiled for 15 min before loaded onto a 4-20%

tris protein gel. The protein gels were run in buffer (25 mM tris, 192 mM glycine, and 0.

1% SDS) at 120 V for 1 hr.

Western blot detection of O-GlcNAc was performed by transferring the proteins

from the gel to a PVDF-membrane in transfer buffer (25 mM tris, 192 mM glycine, and

10% methanol) using a Bio-Rad transfer electrophoresis system at 120 V for 3 hr. Anti-

54 O-GlcNAc MAb CTD 110. 6, an O-GlcNAc specific antibody, in combination with

DyLight 550-conjugate Goat anti-Mouse IgG, a secondary antibody conjugated to a

fluorescent tag, was used to probe the blot for detection of O-GlcNAc. The membranes

were first blocked using 1X BSA blocking buffer (25 mM sodium phosphate,150 mM

NaCl, 0. 1% tween and 1% BSA at pH 7. 4) for 1 hr at room temperature. Anti-O-

GlcNAc antibody and fresh BSA blocking buffer in a 1:5000 dilution were added to the

blot and incubated at room temperature for 1 hr. The blot was rinsed with PBST buffer

(25 mM sodium phosphate, 150 mM NaCl, and 0. 1% tween at pH 7. 4) three times for 5

min each. The secondary antibody was added to fresh 1X BSA blocking buffer and

incubated at room temperature for an additional hour. The blot was rinsed a final time

and imaged using a GE Typhoon 9410 imager. The fluorescent secondary antibody was

excited at 532 nm and detected between 565-595 nm using a 580 nm filter.

3. 2. 8 WGA agarose affinity chromatography

The nickel purified samples were exchanged into WGA wash buffer (20 mM

HEPES, 150 mM NaCl at pH 8) using Amicon centrifugal filter units with a molecular

weight cut-off of 10 kDa to remove the imidazole. The WGA agarose was divided into

100 µL aliquots and washed with 1 mL of wash buffer. The sample was incubated with

the WGA agarose for 20 hr at 4 oC. The mixture was then transferred to a microspin

column and centrifuged at 300 g for 1 min. The agarose was washed with 1 mL of wash

buffer, incubated with 200 µL of WGA elution buffer (20 mM HEPES, 150 mM NaCl, 500 mM GlcNAc at pH 8) for 30 min at 4 oC, and centrifuged again to collect the elution.

55 3. 2. 9 Coomassie protein concentration assay

Nickel-purified and WGA-separated samples were exchanged into WGA wash

buffer using Amicon centrifugal filter units with a molecular weight cut-off of 10 kDa.

Aliquots (10 µL) aliquots were saved for Coomassie analysis. BSA standards were

prepared in WGA wash buffer at 0, 10, 20, 40, 50 μg/mL. Standards and samples were

prepared in 50 μL aliquots. Samples were tested and prepared within the linear range

based on visual inspection. The flow-through and eluents from the WGA purification

were analyzed using a Biotek PowerWave XS microplate spectrophotometer at 595 nm.

3. 2. 10 Supplementing ∆nagZ E. coli cultures with sugar

pCDF-ABL2 and pET43. 1-OGT were transformed into ∆nagZ

Rosetta(DE3)pLysS competent cells and selected on LB agar with spectinomycin,

ampicillin, kanamycin, and chloramphenicol, as previously described. A colony known to

express ABL2 and OGT well was cultured in MM in the presence 0. 5% glucose,

glucosamine, or N-acetylglucosamine in addition to the 0. 4% glucose already in the

MM. At an optical density at 600 nm of 0. 6-0. 8, the cultures were induced with 0. 5 mM

IPTG and grown for 16 hr at 22-25 oC. The cells were harvested using BugBuster

protein extraction reagent and lysed with 3 freeze/thaw cycles and ultrasonication.

Cleared lysate was prepared by centrifugation at 30,000 g for 30 min at 4 oC. Ni

Sepharose 6 Fast Flow, a nickel affinity media, was used to purify the His-tagged ABL2 protein. Gravity flow was used to ensure maximum binding. The nickel affinity media

was equilibrated with wash buffer. The cleared lysate was loaded onto 200 µL of media.

The media was then washed with 2 mL of buffer and the protein was eluted with 1 mL of

300 mM imidazole elution buffer.

56 3. 2. 11 Resolubilized O-GlcNAc-modified ABL2

The insoluble portion of the cell lysates were dissolved in 1 mL of 10% sodium

lauroyl sarcosinate (sarkosyl), followed by dilution to 1% sarkosyl with water and

ultrasonication at 30% for 120 s (3 cycles of 30 s on/10 s off). The resolubilized protein

was centrifuged at 30,000 g for 30 min to pellet any remaining insoluble material and

purified using nickel affinity chromatography as previously described.

3. 2. 12 OGT activity assay

ΔnagZ Rosetta(DE3)pLysS with pCDF-ABL2 and pET43. 1-OGT and ΔnagZ

Rosetta(DE3)pLysS with pCDF-ABL2 with pETDuet-1-codon optimized OGT were expressed for 3 hr at 37 oC. The cells were spun down at 3000 g for 10 min at 4 oC. The cells were resuspended in 50 mM tris and ultrasonicated at 30% for 120 s (3 cycles of

30 s on/10 s off). The lysate was separated from the cells by centrifuging at 30,000 g for

10 min. 1 mM UDP-GlcNAc, 2 μM phenylmethylsulfonyl fluoride (PMSF), 75 units of shrimp alkaline phosphatase, and 12. 5 mM magnesium chloride were added to the lysates and incubated at 20 oC for 1 hr. The mixtures were analyzed by SDS-PAGE and

western blot to determine the relative activity of OGT and codon-optimized OGT. In

addition to O-GlcNAc detection, OGT was detected using Anti-O-GlcNAc Transferase

(DM-17) produced in rabbits and DyLight 650-conjugate Goat anti-Rabbit IgG.

3. 3 Results

3. 3. 1 Expression of O-GlcNAc-modified ABL2 in ΔnagZ E. coli

pET43. 1-OGT and pCDF-ABL2 were co-expressed in ΔnagZ

Rosetta(DE3)pLysS using LB media at 16 oC. Figure 3-3a shows the SDS-PAGE of the

soluble purified ABL2 at a molecular weight of about 14 kDa. Once purified proteins

57 began to precipitate. The precipitated proteins were also analyzed by SDS-PAGE,

Figure 3-3a shows the precipitated protein, referred to as insoluble. ABL2 purified at pH

6 did not produce precipitate or ABL2. Figure 3-3b shows the western blot corresponding to these samples. The O-GlcNAc-modified ABL2 was only in the insoluble sample.

ABL2 was expressed in abundance, but the purification was poor and once purified, it was unstable. The O-GlcNAc-modified ABL2 selectively precipitated with many other proteins. The instability is attributed to purification at high pH (> than 8. 5).

Even though glycosylated ABL2 was temporarily soluble, the amount of O-GlcNAc- modified protein produced was similar to the wild type expressions. ABL2 was purified at a lower pH in an attempt to stabilize the glycoprotein. However, lowering the pH to 6 hindered the lysis and purification of ABL2.

Figure 3-3 O-GlcNAc-modified ABL2 expressed in the ∆nagZ system was unstable and O-GlcNAc-ABL2 selectively precipitated. The soluble and insoluble samples both

58 contain ABL2 as seen on the protein gel (a); however, the western blot (b) shows that O-GlcNAc-modified ABL2 seems to be exclusively in the pellet.

3. 3. 2 ΔnagZ Rosetta(DE3)pLysS sugar trials

pET43. 1-OGT and pCDF-ABL2 were co-expressed in ΔnagZ

Rosetta(DE3)pLysS using minimal media (MM) at 22 oC. Glucose, glucosamine, N-

acetylglucosamine or PUGNAc were added to the MM cultures. The soluble lysate and

the insoluble pellets were analyzed by SDS-PAGE and western blot. Figure 3-4a shows

the SDS-PAGE analysis of cleared lysate, where the asterisk indicates the expression

of ABL2. The corresponding western blot (Figure 3-4b) does not show any glycosylation

for the samples with additional sugar. However, there is a small amount of glycoprotein

in the control sample. Figure 3-4c shows the SDS-PAGE analysis of the insoluble pellet.

The western blot (Figure 3-4d) shows an increased amount of glycoprotein in the glucose and glucosamine supplemented expressions, compared to the control. The

addition of PUGNAc also increased the amount of glycoprotein expressed. However,

ABL2 expressed in the presence of N-acetylglucosamine produced glycoprotein comparable to the control.

Glucose, glucosamine, and N-acetylglucosamine were added to the culture media in an attempt to compensate for the possible decrease of UDP-GlcNAc caused by knocking out nagZ. These monosaccharaides were selected because they participate in the hexosamine biosynthesis pathway. The cultures were grown in MM to control the sugar composition and concentration in the media. The soluble lysate and the insoluble pellets were analyzed by SDS-PAGE and western blot. As previously

observed, little or no glycoprotein was present in the lysate of the cultures (Figure 3-4a).

MM cultures with extra glucose and glucosamine had the highest amount of

59 glycosylation observed thus far. In contrast, the addition of N-acetylglucosamine had no effect on the expression of glycoprotein. The cells with N-acetylglucosamine grew four times faster than the cultures with the other sugars and minimal media alone. The accelerated growth and insubstantial change in glycosylation may be attributed to N- acetylglucosamine’s ability to up-regulate nagA and nagB, which promote glycolysis

[73]. Glycolysis is a process where E. coli uses sugars (mainly glucose) to make ATP, which is required for maintenance and reproduction of the cells [74]. The addition of

PUGNAc also increased the amount of glycosylation, suggesting the presence of a second glycosidase with activity toward O-GlcNAc-modified proteins in E. coli.

Figure 3-4 Analysis of sugar supplemented expression of O-GlcNAc modified ABL2. The cleared lysate from cultures supplemented with glucose have a small amount of ABL2 (a) that is not glycosylated (b). The pellets corresponding to these samples contain about the same amount of ABL2 (c); however, the samples supplemented with glucose and glucosamine have an abundance of O-GlcNAc-modified ABL2 (d). The pellets containing PUGNAc express more ABL2 (c), but less O-GlcNAc-modified ABL2 compared to the glucose and glucosamine samples (d).

3. 3. 3 Resolubilized ABL2

The insoluble glycoprotein from Figure 3-4d was solubilized using detergent and isolated from insoluble content and cell debris using nickel affinity chromatography.

61 Figure 3-5a shows the SDS-PAGE analysis of purified resolubilized ABL2. There were

very small amounts of ABL2 and some other proteins observed. The western blot

(Figure 3-5b) shows that these results were similar to previous results. The glucose,

glucosamine, PUGNAc, and PUGNAC+glucose samples had higher levels of

glycosylation than the control and N-acetylglucosamine sample.

The overall goal of this project is to express milligram quantities of soluble O-

GlcNAc-modified protein for NMR analysis. In an effort to solubilize O-GlcNAc-modified

ABL2 without denaturing the protein, sarkosyl was used to solubilize any protein from the inclusion bodies. The results were similar to the previous experiments: glucose and glucosamine appeared to have more glycoprotein than both the control and the N- acetylglucosamine sample. The glucose in combination with PUGNAc appeared to have the highest level of purified O-GlcNAc-modified ABL2. The asterisks in Figure 3-5b indicate multiple O-GlcNAc-modified ABL2 bands that migrate differently in the SDS-

PAGE gel. More than one O-GlcNAc modification can occur on ABL2, so the bands may correspond to different numbers of O-GlcNAc residues. There are a few proteins that

consistently co-purify with ABL2, suggesting that O-GlcNAc-modified ABL2 may be bound to or tightly associated with these proteins.

Figure 3-5 Purification of resolubilized O-GlcNAc-modified ABL2. The pellets were resolublized with 10% sarkosyl, diluted to 1%, and purified with nickel affinity chromatography. The results were very similar to previous data—there is not an abundance of ABL2 on the protein gel, (a) yet it is highly glycosylated (b).

3. 3. 4 Optimized expression of O-GlcNAc-modified ABL2 in ∆nagZ

Previous experiments produced highly glycosylated protein (Figure 3-4d) that was insoluble. This glycoprotein can be isolated from the pellet using a sarkosyl.

Although the detergent experiments worked, it is not clear whether sarkosyl will affect

the protein’s structure. In an effort to optimize the expression, for the production of

soluble O-GlcNAc-modified ABL2, the cells were cultured in different conditions. There

are many variables that can greatly affect the expression of soluble target protein. The

cells were cultured in a tissue culture flask at slower revolutions per minute (rpm). The

flat tissue flask doubled the surface area of aeration. These expressions were done at

room temperature (22-25 oC) instead of a controlled temperature of 22 oC for 18 hrs.

The expressions were induced with IPTG at an OD600 of 1. 0, instead of 0. 6. Most

importantly, 10 μM PUGNAc was added to all of the purification buffers. The samples

63 were purified and concentrated for analysis. The sample with 0. 5% glucose was the

only sample that showed significant amounts of pure, soluble O-GlcNAc-modified protein (Figure 3-6). Although the overall expression of ABL2 appears to be low, the ratio of glycoprotein to naked protein was high. In contrast to expression in wild type E. coli, O-GlcNAc-modified ABL2 expressed in ∆nagZ E. coli remained glycosylated in the presence of PUGNAc after several days. O-GlcNAc-modified ABL2 expressed in the wild type, in the presence of 10 μM PUGNAc lost the modification after 2 days. It is not clear whether the decreased amount of ABL2 is real or if the glycosylation disrupts the binding of the coomassie G-250 dye, used for detection of protein. Coomassie is also the reagent used in protein concentration assays, and is not compatible with high levels of GlcNAc.

Figure 3-6 Optimized expression of soluble O-GlcNAc-modified protein. A starter culture from a single colony was used to inoculate sugar-supplemented cultures in MM. The expression conditions were optimized to produce soluble O-GlcNAc-modified protein. The soluble lysate was purified and concentrated for SDS-PAGE and western

64 blot analysis. The glucose sample was the only sample to produce a substantial amount of soluble O-GlcNAc-modified protein.

3. 3. 5 Comparison of OGT activity produced from human OGT (hOGT) and codon-optimized OGT (coOGT)

The human OGT sequence contains 52 rare codons. Rare codons are DNA

codons that have a low amount of corresponding tRNA, required for protein synthesis,

in E. coli. The OGT sequence was codon optimized, cloned into pETDuet-1 by the

protein facility at LSU, and overexpress in E. coli. Although codon optimized OGT over-

expressed in ∆nagZ E. coli, it did not show any glycosylation, even when glucose and glucosamine were added as supplements to the media. To verify that the coOGT was

active, cleared lysate containing coOGT was used to glycosylate ABL2 in vitro, and the

relative amount of glycosylation was compared to hOGT in vitro glycosylation. Figure 3-

7 shows the SDS-PAGE and western blot of the in vitro glycosylation reaction. There is

more coOGT than hOGT and there is more O-GlcNAc-modified protein produced from

the hOGT reaction. A comparison of the relative band densities of the western blot

bands to the corresponding SDS-PAGE band for the in vitro coOGT and hOGT

reactions show that the extent of glycosylation is similar for both enzyme. The same

cleared lysate was analyzed to assess the relative concentrations of OGT. coOGT was

more abundant than hOGT. In theory more coOGT should produce more O-GlcNAc-

modified protein, but it does not. These results suggest that coOGT is not as active as

hOGT, likely because it does not fold properly during translation.

Figure 3-7 Comparison of human OGT and codon optimized OGT. Anti-O-GlcNAc transferase was used to detect the amount of OGT present in the in vitro reaction. Anti- O-GlcNAc was used to detect the amount of O-GlcNAc-modified protein produced from the reaction.

3. 4 Conclusion

The ΔnagZ Rosetta(DE3)pLysS E. coli co-expression system produced more O-

GlcNAc-modified ABL2 compared to wild type Rosetta(DE3)pLysS E. coli, but only

when the media was supplemented with extra glucose or glucosamine. The insolubility

of O-GlcNAc-modified ABL2 was a problem for both E. coli strains, providing more

evidence that co-translational modification with O-GlcNAc disrupts ABL2 folding.

Another possibility is that a second glycosidase is removing O-GlcNAc from soluble

ABL2. Evidence for this hypothesis includes increased levels of O-GlcNAc-modified

ABL2 when PUGNAc is added to cultures of ΔnagZ Rosetta(DE3)pLysS E. coli co-

expressing OGT and ABL2. Since O-GlcNAc-modified ABL2 is available in the pellet,

resolubilized O-GlcNAc-modified ABL2 may be available in sufficient quantities for NMR

analysis. Another possible solution to producing O-GlcNAc-modified protein in sufficient

quantities for NMR analysis is to switch the target protein to one that has been modified

66 with in vivo methods. David Vocadlo has had great success with tau and it is biologically relevant. [24]

In vitro glycosylation could be a viable method to produce mg quantities of O-

GlcNAc-modified protein. In the past, in vitro glycosylation has not been efficient, but

NagZ may have contributed to this apparent inefficiency. The ∆nagZ E. coli may allow for in vitro glycosylation without the need for β-N-acetylglucosaminidase inhibitors.

Optimizing the expression of human OGT is also very important. Increasing the amount of hOGT expressed, while maintaining its activity, should increase the amount of protein being glycosylated in the lysate. Partial codon optimization may lead a properly folded, active, and abundant OGT. Rosetta cells contain tRNA for 6 common rare codons.

There are 52 rare codons in the hOGT sequence, Rosetta compensates for all but 10.

These 10 codons are all CGA. Mutating these codons, without changing the amino acid may lead to more protein expression.

Wheat germ agglutinin is often used to enrich O-GlcNAc-modified proteins, but it is expensive and is not always efficient because of its weak affinity for terminal GlcNAc.

In a situation where there is only a very small percent of O-GlcNAc-modified protein, enrichment is important for down-stream analysis like mass spectrometry and NMR.

Galactosyltransferase adds a galactose to O-GlcNAc. This enzymatic process has been used to detect O-GlcNAc. While the addition of functionalized galactose provides other ways to analyze the modification. The functionalized galactose has a reactive group that can be used to bind biotin for enrichment with streptavidin [75]. The addition of a mass tag, such as polyethylene glycol (PEG) has been used to visualize multiple-O-GlcNAc-

67 modified proteins by SDS-PAGE gel and could be used to give valuable information such as how much of the protein is O-GlcNAc-modified and to what degree [33].

While there is still some optimization required for the co-expression system to produce mg quantities of soluble O-GlcNAc-modified protein, there are some promising opportunities that need to be explored. The resolubilized samples are highly glycosylated and the structures of these proteins should be analyzed. If they are properly folded, they may lead to an NMR structure. Even with small amounts of resolubilized O-GlcNAc-modified protein, mass spectrometry can be used to map the O-

GlcNAc sites to specific residues. This information alone can provide valuable insight into the role of O-GlcNAc in the structure and function of proteins.

68 Chapter 4: Chlamydia trachomatis CT663: Structure and Ligand Interaction Model

4. 1 Introduction

4. 1. 1 Chlamydia trachomatis

Chlamydia is the most common sexually transmitted bacterial disease in the

United States [76]. There are 2. 8 million new infections annually [77]. Chlamydia can cause reproductive health complications such as pelvic inflammatory disease, infertility and ectopic pregnancy. It can also cause neonatal conjunctivitis and pneumonia [78].

The disease is asymptotic in 50-80% of patients, so they are often unaware of the infection [79]. The rate of re-infection is alarmingly high [80]. Although Chlamydia is both curable and preventable, there is a need for better therapies. In 2000, three chlamydial strains in clinical patients demonstrated multi-drug resistance to doxycycline, azithromycin, and ofloxacin [79].

Chlamydia is an obligate intracellular bacteria, which means it requires a host cell for replication. It infects columnar epithelial cells, which are found in the uterus, the respiratory tract, and the eye. The disease manifests in these areas as a urogenital infection, pneumonia, and conjunctivitis, respectively. The developmental cycle is unique and consists of two alternating cellular forms. The infectious form, elementary bodies (EB), attach to host cells and initiate the developmental cycle, after endocytosis

[81]. The EBs develop into the noninfectious replication form called reticulate bodies

(RB). These RBs reorganize in an asynchronous manner and then revert back to EBs.

The now multiplied EBs exit the host cells, lysing the cells in the process [82]. The newly produced EBs infect new cells and the cycle starts again.

Figure 4-1 Developmental cycle of Chlamydia. Brunham, R. C. ; Rey-Ladino, J. , "Immunology of Chlamydia infection: Implications for a Chlamydia trachomatis vaccine," Nature Reviews Immunology 2005, 5 (2), 149-161.

4. 1. 2 Type III secretion

The elementary bodies use the virulence associated type III secretion system

(T3SS) to infect the host cells. The T3SS is used by gram-negative bacterial pathogens to deliver virulence factors, molecules essential for the induction of infection, into the

cytosol of host cells [3]. Virulence factors can be a wide variety of molecules, and are

usually on the bacterial surface or secreted from the pathogen. Because the molecules

are critical for the initiation and replication of the disease, they are sensible targets for

drug development [83]. The T3SS is composed of an injectisome, an ATPase,

translocator molecules, effector molecules, and chaperones [84]. The injectisome is the

major machinery of the virulence T3SS. It is a multi-protein complex that consists of a

70 basal body embedded within the bacterial membrane topped by a needle-like structure

[85]. It serves as a channel through which proteins reach the host cell. Proteins are secreted through the injectisome by an ATP consuming process, using ATPases located in the bacterial cytosol [84]. Translocator proteins are secreted to form the translocon, which sits at the needle-like end of the injectisome. The translocon physically links the injectisome to the host cell, enabling secretion of virulence factors into the host cell.

Effectors are eukaryotic-like proteins that modulate the host cell throughout the infectious cycle. Bacterial pathogens secrete effectors that manipulate the actin cytoskeleton of the host cell, which enables remodeling of the membrane in order to promote uptake or adhesion of the bacteria [84]. Effectors and translocators are often bound to chaperones in the bacterial cytosol. Chaperones are proteins that bind to virulence factors in order to keep them in secretion-competent state as they travel from the bacteria and through the injectisome. The chaperones also prevent aggregation and premature interaction of virulence factors, as well as regulate transcription of T3SS

71 genes [87]. T3SS chaperones can be categorized into three classes based on structure and substrate similarity. Class I chaperones are approximately 130 residues. Class IA chaperones associate with only one effector, while those in class IB associate with multiple effectors. Both sub-classes of class I adopt homo-dimeric structural folds. Class

II chaperones are approximately 160 residues and associate with translocator proteins.

They are structurally distinct, typically containing three tetratricopeptide repeats (TPR)

[84, 88]. Class III chaperones are extra-cellular filament proteins that have extended

alpha helical structure.

4. 1. 3 CT663/Scc1-CopN complex

CT663 is a class 1A chaperone found in Chlamydia trachomatis. It is a dual

function enzyme; it functions as a chaperone and as a transcription factor that regulates

the global transition from RB to EB in the Chlamydial developmental cycle. CT663

carries out this function by associating with the β subunit (RpoB) and the primary sigma

subunit σ66 (RpoD) of Chlamydial RNA polymerase (RNAP). This association inhibits σ66

dependent transcription [89]. Analysis of CT663 as a chaperone has shown that it

interacts with secretion substrates as well as another class IA T3SS chaperone, Scc1

[84]. This interaction suggests that CT663 and Scc1 form a heterodimeric chaperone

complex. The Ct663/Scc1 complex binds Chlamydial outer protein N (CopN), a T3SS

effector that is essential for virulence. CopN destroys microtubules in cytoskeleton

networks, resulting in cell cycle arrest. CopN is also a T3SS plug protein [90]. Plug

proteins typically block the secretion pore, allowing protein export only in the presence

of a currently unknown extracellular signal [84]. Although the formation of a CT663,

72 Scc1, and CopN complex has been confirmed, the order of binding events and

arrangement are unknown.

4. 1. 4 Research goal

The goal of this research is to determine the structure and analyze the ligand

interactions of CT663. Analysis of CT663’s interactions with σ66, β-subunit, Scc1, and

CopN will provide information about how CT663 functions as a dual enzyme. Structural

analysis of CT663 and its binding partners will provide a rational drug model.

Knowledge of how these proteins interact will facilitate the design of a small molecule

that inhibits key interactions necessary for infection. An x-ray crystallographic structure

of CT663 has been attempted by David Worthylake, a collaborator at LSU Health

Science Center School of Medicine, but crystal formation has not been successful.

Nuclear magnetic resonance (NMR) spectroscopy will be used to determine the

structure of the chaperones CT663 and Scc1, as well as the chaperone complex

CT663/Scc1. Structural determination of any of the chaperones bound to the effector

CopN is not feasible, due to size constraints. NMR can still provide information about

these interactions, such as which residues participate in binding and how binding affects

the structure of the chaperones. Further binding analysis will be done using surface

plasmon resonance (SPR) spectroscopy to determine the binding kinetics and

arrangement of CT663, Scc1, and CopN.

73 4. 2 Materials and Methods

4. 2. 1 Materials

Glycerol, glucose, sodium chloride (NaCl), sodium phosphate dibasic, sodium

phosphate monobasic, 2-(N-morpholino)ethanesulfonic acid (MES), ammonium acetate,

dithiothreitol (DTT), ethylenediaminetetraacetic acid (EDTA), bromophenol blue,

Coomassie brilliant blue R250, sodium dodecyl sulfate (SDS), glycine,

tris(hydroxymethyl)aminomethane (tris), 2-mercaptoethanol, imidazole, urea, acetic

acid, methanol, kanamycin, ampicillin, and chloramphenicol were purchased from

Sigma-Aldrich. Rosetta(DE3)pLysS cells, BugBuster protein extraction reagent, 3K and

10K Amicon centrifugal filters, and ampicillin were purchased from EMD

Millipore/Novagen. Precision Plus Protein Unstained Standards, 4–20% mini-PROTEAN

TGX precast gels, and Laemmli buffer were purchased from Bio-Rad. BioReagents glutathione agarose, HRV 3C protease, EZ-Run Rec protein ladder, and Coomassie

(Bradford) protein assay kit were purchased from Thermo Scientific/Pierce Antibodies.

Luria broth (LB) and agar were purchased from BD Biosciences. Hiprep 16/60

Sephacryl-S-100 HR and Ni Sepharose 6 Fast flow were purchased from GE

Healthcare. Other chemicals and materials used include isopropyl β-D-1-

thiogalactopyranoside (IPTG, Affymetrix Anatrace), deuterium oxide (D2O, Cambridge

Isotopes), tween 20 (Fisher), chloramphenicol (calbiochem), QIAprep spin miniprep kit

(Qiagen), Costar 96 well UV transparent flat bottom plates were also purchased.

74 4. 2. 2 Structure predictions

The amino acid sequence of CT663, Scc1, and CopN were submitted to an online protein homology/analogy recognition engine version 2. 0 (Phyre2) for structure

prediction.

4. 2. 3 CT663 expression and purification

pET21-CT663 was transformed into Rosetta (DE3)pLysS competent E. coli cells

and plated on Luria broth (LB) agar with 100 µg/mL ampicillin and 35 µg/mL

chloramphenicol. Single colonies were selected for inoculating 5 mL of LB as starter

cultures and grown overnight at 37 oC with a shaking speed of 250 revolutions per

minute (rpm). The starter culture was added to 0. 5 L of LB media and grown at 37 oC

with 250 rpm. Once the culture reached an OD600 of 0. 6-0. 8, it was induced with 0. 5

mM IPTG. The culture grew at 37 oC for 3 hr and the cells were harvested at 4,000 g for

10 minutes at 4 oC. The cells were reconstituted in 10 mL of BugBuster followed by

ultrasonication at 30% power for 120 s (3 cycles of 30 s on/10 s off) using a Sonics

Vibra-cell VCX130. Cleared lysate was prepared by centrifugation at 30,000 g for 30

min at 4 oC. CT663 was purified from the cleared lysate using 1 mL of Ni Sepharose 6

fast flow media by gravity flow. CT663 was washed with 20 mL of 20 mM sodium

phosphate, 300 mM NaCl, 20 mM imidazole, pH 8 buffer and eluted in 10 mL of buffer

with 300 mM imidazole.

4. 2. 4 CT663 and Scc1 co-expression and co-purification:

E. coli cells containing pACYC184-Scc1-6XHis and pET28-CT663 were sent to

us by our collaborator, Dr. Li Shen, at the LSU Health Science Center School of

Medicine. The cells were plated on LB agar with 35 µg/mL chloramphenicol and 50

75 µg/mL kanamycin. 5 mL starter cultures were grown from single colonies. The starter culture was added to 0. 5 L of LB media at 37 oC. Once the culture reached an optical density of 0. 6-0. 8 at 600 nm, they were induced with 0. 5 mM IPTG and 0. 2% arabinose. The cells were harvested and the Scc1:CT663 complex was co-purified following the same procedures used for CT663 (see section 4. 2. 1).

4. 2. 5 CopN expression and purification

E. coli cells containing pGEX098-CopN were sent to us by our collaborator, Dr. Li

Shen, at the LSU Health Science Center School of Medicine. The cells were plated on

LB agar with 100 µg/mL ampicillin. The cells were grown and harvested following the same procedures used for CT663 (see section 4. 2. 2). CopN was purified using 1 mL of glutathione agarose media by gravity flow. The media was equilibrated with 10 mL of 50 mM tris, 150 mM NaCl, pH 8 buffer (TBS). The cleared lysate was added to the media and gently mixed at 4 oC for 30 min. The media was washed with 20 mL of TBS. The

GST-fusion tag was cleaved from CopN using 20 units of HRV 3C protease in 2 mL of

TBS with gently mixing at 4 oC for 16 hr. The flow-through containing CopN and HRV 3C was collected, and HRV 3C was removed by passing the sample over 0. 5 mL of nickel affinity media.

4. 2. 6 Separation of CT663 and Scc1

Previous research indicates that Scc1 is insoluble when expressed alone.

However, when it is co-expressed with CT663, it forms a soluble complex. The

CT663/Scc1 complex was expressed and purified as previously described. 6M urea, 8M urea, 2% SDS and 1% sarkosyl were used to separate the complex. Fast protein liquid chromatography (FPLC) was used to bind the complex to nickel media. The media was

76 washed with wash buffer containing one of the separating agents. Any remaining bound

protein was eluted with 20 mM sodium phosphate, 300 mM NaCl, and 300 mM

imidazole at pH 8 buffer. The UV absorbance was monitored for the release of protein

during the separation wash. SDS-PAGE was used to analyze the proteins that eluted

after separation.

4. 2. 7 Gel electrophoresis analysis

Purified CT663, Scc1/CT663 complex, and CopN were run in 4-20% tris-HCl gels under denatured and native conditions. The denatured samples were prepare in a 1:1

ratio with Lammeli buffer containing 5% v/v 2-mercaptoethanol and heated at 95oC for

15 minutes. The samples were ran at 125 V for 60 min in 25 mM Tris, 192 mM glycine,

1. 0% SDS, pH 8. 3 running buffer. Prior to running the native samples the 4-20% tris-

HCl gel was run in 25 mM tris, 192 mM glycine, pH 8. 3 native running buffer for 60 min to remove the SDS from the gel. Samples prepared for the native gel were buffer exchanged into 20 mM tris, 100 mM NaCl, pH 8 buffer and concentrated to 25 µM in 20

µL using Amicon centrifugal filter units with a molecular weight cut-off of 10 kDa. Each

sample was run individually along with 1:1 mixtures of CopN:CT663 and

CopN:Scc1/CT663. These mixtures were incubated at 25oC for 1. 5 hr. 10% glycerol

was added to the native sample in a 1:2 ratio and the native gel was fun for 75 min at

125 V.

4. 2. 8 Gel filtration analysis

CT663 and the Scc1/CT663 complex were analyzed with quantitative gel filtration analysis using either a HiPrep Sephacryl S-100 HR with a 1 mL sample injection volume or an ENrich 70 SEC column with a 50 µL sample injection volume. Two stock solutions

77 of CT663 and two stock solutions of the Scc1/CT663 complex were prepared for

analysis using Amicon centrifugal filter units with a molecular weight cut-off of 10 kDa.

The concentrations and buffer conditions of each sample are listed in Table 4-1. Protein

concentrations were determined using the Bradford colorimetric assay with Coomassie

Plus reagent and a Biotek PowerWave XS microplate spectrophotometer. Standard

curves were generated using molecular weight standards (2 mg/mL BSA, 1 mg/mL β- lactoglobulin, 1 mg/mL cytochrome C, and 50 mM cytidine) using each running buffer:

(1) 20 mM tris, 150 mM NaCl, pH 8. 0, (2) 20 mM tris, pH 7. 4, (3) 50 mM potassium phosphate monobasic/sodium phosphate dibasic (K-Na-phosphate), pH 7. 3, and (4) 20 mM sodium acetate, pH 4. 5. Regression analysis was used to determine the linear relationship between the elution volume and the logarithm of the molecular weight for each column/buffer combination. The molecular weights of the samples were determined using the linear function to determine the oligomeric state of the samples in each buffer.

4. 2. 9 SPR Analysis

Binding experiments were conducted using a kinetic titration method called

single-cycle kinetics. This method consists of serial injections of analyte at increasing concentrations over the surface of a ligand-functionalized sensor chip surface without regeneration between sample injections [91]. SPR studies were performed at 25 oC on

a nitrilotriacetic acid (NTA) chip (Xantec) on a BIAcore T100 system (GE Healthcare).

Both cells of the uncharged NTA chip were conditioned with 0. 5 M EDTA at a flow rate

of 10 μL/min with 180 s of contact time, followed by 0. 1 M NaOH at a flow rate of 10

μL/min with 180 s of contact time. Two capture cycles were performed on the second

78 cell of the sensor chip to create a capture baseline. The capture cycle included three

steps: charging the surface with Ni2+, capturing the ligand, a series of washes, and

surface regeneration. The second cell of the chip was charged with 5 mM NiCl2 at a flow rate of 10 μL/min with a contact time of 120 s. Charging the surface with nickel allows the binding of histidine tagged ligands, such as the His-Scc1/Ct663 complex. The His-

Scc1/CT663 complex was prepared in running buffer (50 mM potassium phosphate,

150 mM NaCl, 5 mM EDTA, 0. 005 % Tween at pH 7. 3) and immobilized onto the

surface at a flow rate of 10 μL/min with a contact time of 150 s. The chip was

regenerated with 0. 5 M NaOH at a flow rate of 30 μL/min for 2 min, followed by 350 mM

EDTA at 30 μL/min for 2 min. The capture cycle was repeated. The analyte binding step

was a single cycle kinetic method with 5 increasing concentrations. This cycle consisted

of a charging step, a ligand capture step, followed by 5 analyte binding and dissociation

steps. CopN, the analyte, was flown at a rate of 10 μL/min with 150 s of contact time

across the second cell with a dissociation time of 240 s. The parameters of the other

steps were the same. CopN was passed across the surface of the first flow cell as a

control or blank.

4. 2. 10 NMR buffer optimization for CT663

CT663 was expressed in minimal media as previously described (Chapter 3).

15 Isotopic labeling was incorporated by adding NH4Cl to the MM. The proteins were

purified with Ni-affinity chromatography as previously described. Purified CT663 was

purified with size exclusion chromatography using a Hiprep 16/60 Sephacryl-S-100 HR

column in 20 mM PBS at pH 7. 15N-CT663 was concentrated to 500 µM and exchanged into 20 mM sodium phosphate, 20 mM tris, and 20 mM MES buffers using Amicon

79 centrifugal filters units with a molecular weight cut-off of 3 kDa. The samples were

concentrated to a final volume of 250 µL and transferred to 5 mm Shigemi tubes with

1 15 10% D2O. An H - N HSQC spectrum of each sample was collected on a Varian 700

MHz NMR.

4. 2. 11 15N-CT663/Scc1 NMR analysis

15N-CT663 was expressed, purified, exchanged into 20 mM sodium acetate at pH

6 buffer, and concentrated to 500 µM as previously described. pACYC-184-Scc1 was

transformed into BL21 cells and selected on LB agar with 35 µg/mL chloramphenicol. A

single colony was added to 10 mL of LB and grown overnight. This starter culture was

o added to a 1L of LB and cultured at 37 C until it reached an OD600 of 0.6-0.8. The cells

were induced with 0. 5 mM IPTG and cultured at 37 oC for 3 hr. The cells were

harvested by centrifugation at 4000 g for 10 minutes at 4 oC and reconstituted in 10 mL

of 8 M urea in BugBuster. Sonication at 30% power pulsed for 120 s (3 cycles of 30 s

on/10 s off) was used to lyse the cells, and cleared lysate was prepared by

centrifugation at 30,000 g for 30 min at 4 oC. After filtering the cleared lysate with a 0.

22 µM syringe filter. The denatured Scc1 was bound to Sepharose 6 fast flow media

using the FLPC. The media was washed with 10 mL of 20 mM sodium phosphate, 300

mM NaCl, 20 mM imidazole, and 6 M urea at pH 8 buffer. A 30 mL linear gradient at 0.

1 mL/min from 6 M urea to 0 M urea was used to refold the protein on the column.

Refolded Scc1 was eluted with 20 mM sodium phosphate, 300 mM NaCl, and 300 mM

imidazole at pH 8 buffer. Scc1 was exchanged into 20 mM sodium acetate at pH 6

buffer and added to the 15N-CT663 sample. The 1H -15N HSQC spectra of 250 µL of 500

80 µM 15N-CT663 before and after the addition of Scc1 were collected on a Varian 700

MHz NMR.

4. 3 Results

4. 3. 1 Structural predictions of the Chlamydial trachomatis T3SS effectors- chaperone complex CT663:SCC1:CopN

The T3SS is not exclusive to Chlamydia trachomatis and is used by other pathogens as well. The T3SS apparatus itself has been shown to display notable similarities among different bacterial species [92]. E. coli, Yersinia, Salmonella, and

Shigella are a few examples of pathogens that use the T3SS. The macromolecules of the T3SS are well characterized for these pathogens. High-resolution crystal structures of all classes of T3SS chaperones, crystal structures of the proteins involved in basal body formation, and NMR and FTIR analyses of needle proteins have advanced the structural knowledge of the T3SS [92]. Using the structures of chaperones and effectors from other bacterial species as a basis has allowed for a structural prediction of

CT663/Scc1 in complex with CopN. Protein homology/analogy recognition engine version 2. 0 (Phyre2) is a structure prediction server that uses homology/analogy of known structures to predict the structure of other proteins. Structural homology is based on two principles: “first, the structure of a protein is determined by it’s amino acid sequence; second, proteins with similar sequences are practically identical in structure, and even proteins that have distantly related sequences still fold into similar structures”

[93]. Protein analogy refers to proteins with similar functions despite their different sequences. The structure of CT663 was predicted with 97% sequence coverage from 4 template structures of T3SS chaperones from other bacteria [94-97]. The structure of

81 Scc1 was predicted based on three proteins with 92% coverage [94, 97, 98]. The structural orientation of CopN in complex with CT663/Scc1 was predicted based on a

Yeseria pestis effector chaperone complex (PDB 1XKP) [99]. The Yeseria pestis complex as well as a protein from Shigella flexneri was used to predict the structure of

CopN with 92% coverage [99, 100]. The structure prediction was done to better understand how CT663 forms a complex with SCC1 and CopN.

Figure 4-3 The CopN:CT663:Scc1 structure is modeled using Yersinia pestis virulence factor YopN in complex with the heterodimeric chaperone SycN-YscB by Phyre2.

4. 3. 2 Expression and Purification

The expression and purification conditions were optimized by Dr. Li Shen, our collaborator at the LSU Health Science Center School of Medicine. CT663, Scc1/Ct663, and CopN express with yields of ~5 mg, ~66 mg, and ~2 mg of protein per liter of culture, respectively. The purification of these proteins is relatively clean. Figure 4-4 shows the SDS-PAGE and native gel analysis of the purified proteins. The SDS-PAGE gels separate denatured proteins based on their size and charge. While native gels separate the proteins in the same manner, the proteins are in their native state and SDS

82 is not available to bind and add negative charge. Without SDS to denature and add

uniform negative charge, it is hard to determine the exact mass of the native proteins.

However, it appears that CT663 and CT663/Scc1 form dimers under native conditions.

CopN was cloned and expressed with a glutathione S-transferase (GST) fusion tag for

purification. This GST tag forms dimers, so the CopN dimer was expected and not

informative. While Figure 4-4a shows CopN cleaved from GST, the CopN used for the

native gel is not cleaved. A quick and dirty binding assay, mixing proteins in 1:1 ratio,

showed no binding of the chaperones with CopN. This result is inconsistent with gel

filtration analysis and is likely an artifact from interactions with the gel and/or

interference from the GST fusion tag.

Figure 4-4 SDS-PAGE and native PAGE analysis of CT663, Scc1/CT663, and CopN. The protein molecular weight ladder is shown in lane 1 in the (a) SDS-PAGE and (b) native PAGE gel images. The molecular weight of CT663, Scc1/Ct663 and CopN are 14. 7 kDa, 30. 9 kDa, and 45. 2 kDa respectively.

4. 3. 3 Separation of CT663/Scc1 complex

The CT663/Scc1 complex is expressed within E. coli as a complex and also co- purifies. The fact that CT663 and Scc1 purify together suggests a very low dissociation

83 constant. Since Scc1 is difficult to produce when expressed in E. coli alone, we tried to

separate CT663 and Scc1 and recover each protein in a soluble, folded state. The

complex was bound to nickel affinity media through the His-tag on Scc1. Urea was used to try to separate and elute CT663 from Scc1. 6 and 8 M urea were not sufficient to destabilize the interaction between the CT663 and Scc1. Urea is a very common compound used to denature proteins. The exact mechanism of how urea denatures proteins is not known, but studies have shown that urea disrupts hydrophobic

interactions. Urea also interacts with polar residues and the peptide backbone to

stabilize the denatured form of the protein [101]. The fact that 8 M urea is not enough to

separate CT663 from Scc1 suggests that interactions other than hydrophobic, like

electrostatic and hydrogen bonds, contribute to the stability of the complex.

Figure 4-5 CT663-Scc1 interaction. Purified CT663 and CT663:Scc1 complex are shown in lanes 2 and 3, respectively. The purified CT663-Scc1 complex was immobilized on 1 mL of Ni-affinity media via a 6XHis tag on Scc1 and washed with 20 mL of wash buffer with 6 M urea (lane 4), 8 M urea (lane 5), and 2% SDS (lane 6). Even under strong denaturing conditions of 8 M urea, some CT663 remained bound to 6XHis- tagged Scc1. SDS (2%) was sufficient to disrupt the CT663:Scc1 interaction, resulting in pure Scc1.

84 4. 3. 4 Buffer dependent gel filtration analysis

Gel filtration was done in an effort to characterize the oligomeric state of the chaperones CT663 and CT663/Scc1. The goal of this study was to find conditions where the chaperones were in their monomeric state. CT663 and Scc1 formed a heterodimer in 20 mM sodium acetate at pH 4. 5, while CT663 formed high order oligomers regardless of buffer conditions.

Figure 4-6 Gel filtration analysis of CT663 and CT663:Scc1 at various buffer conditions. Two gel filtration column were used to analyze the samples: (A) a HiPrep Sephacryl S- 100 HR column using a 1 mL sample volume and (B) an Enrich SEC 70 column using 50 L sample volumes. Molecular weight standards were used to determine the linear relationship between the elution volume and the logarithm of the molecular weight for each column/buffer combination. The molecular weight of the samples were determined

85 using the best-fit linear function based on regression analysis to determine the oligomeric state of the samples in each buffer. The sample concentrations were 81-152 µM except for the CT663:Scc1 complex in sodium acetate, which was 27 µM due to its instability in this buffer. Details about the sample conditions are described in Table 4-1. (A) In Tris/saline buffer at pH 8. 0, CT663 (solid black line) predominantly exists as higher order oligomers (>300 kDa) and as a homodimer while the CT663:Scc1 complex (dashed black line) predominantly forms a stable tetramer. In acetate buffer at pH 4. 5, CT663 and the CT663:Scc1 complex are unstable. CT663 readily precipitates, but the complex stability was better and showed that soluble CT663:Scc1 in acetate buffer (gray solid line) exists as a heterodimer. (B) In phosphate buffer at pH 7. 3 (solid black line), CT663 forms a homodimer and in Tris buffer at pH 7. 4 (dashed black line), CT663 forms homotrimers.

4. 3. 5 Detergent assisted Gel filtration

Since different buffers and salt did not separate CT663 oligomers, detergents

were added in an attempt to stabilize the monomeric form of CT663. When the

detergents were added to Tris buffer, it increased the aggregation of the protein (Figure

4-6a). The detergents had little effect on CT663 in phosphate buffer, as shown in Figure

4-6b.

Figure 4-7 Detergent effects on CT663 oligomerization. (A) CT663 samples were run in 20 mM Tris, pH 7. 4 buffer (solid black line) with 1 mM SDS (dotted black line), 10 mM SDS (dashed black line), and 0. 01% Triton X-100 (dash-dotted black line). The predicted elution volumes for CT663 oligomers (+) were calculated using the calibration curve (Table 4. 1). The number of monomer units for each oligomer is shown above the marker. (B) CT663 samples were run in 50 mM potassium-sodium phosphate, pH 7. 3 buffer (solid gray line) with 0. 01% Triton X-100 (dotted gray line) and 1% PEG (dashed gray line). The predicted elution volumes for CT663 oligomers (+) are shown as in (A). Each sample was 50 L of 81 M CT663 in 50 mM K-Na phosphate, pH 7. 3 buffer, except for the sample run in 10 mM SDS/Tris buffer. This sample was diluted by a factor of 4 with the running buffer to avoid SDS-phosphate precipitation. The absorbance values for this sample were multiplied by 4 for direct comparison to the other samples.

Table 4-1 This table summarized the resulting oligomeric state of CT663 and Scc1/CT663 complex in different buffer conditions.

Sample Column/Sample (uM) Sample Buffer Running Buffer volume Results 20 mM sodium phosphate, HiPrep CT663 20 mM Tris, 150 mM hexamer, 300 mM imidazole, 300 mM Sephacryl S-100 (152) NaCl, pH 8. 0 dimer NaCl, pH 8. 0 HR/ 1 mL 20 mM sodium phosphate, HiPrep Scc1:CT663 20 mM Tris, 150 mM hetero- 300 mM imidazole, 300 mM Sephacryl S-100 (133) NaCl, pH 8. 0 tetramer NaCl, pH 8. 0 HR/ 1 mL HiPrep Scc1:CT663 20 mM sodium acetate, pH 20 mM sodium hetero- Sephacryl S-100 (27) 4. 5 acetate, pH 4. 5 dimer HR/ 1 mL 50 mM K-Na-phosphate, pH Enrich 70 SEC/ CT663 (81) 20 mM Tris, pH 7. 4 trimer 7. 3 50 uL 50 mM K-Na-phosphate, pH 50 mM K-Na- Enrich 70 SEC/ CT663 (81) dimer 7. 3 phosphate, pH 7. 3 50 uL 50 mM K-Na-phosphate, pH 1 mM SDS, 20 mM Enrich 70 SEC/ apparent CT663 (81) 7. 3 Tris, pH 7. 4 50 uL pentamer 50 mM K-Na-phosphate, pH 10 mM SDS, 20 mM Enrich 70 SEC/ apparent CT663 (20) 7. 3 Tris, pH 7. 4 50 uL pentamer

87 50 mM K-Na-phosphate, pH 0. 01% Triton X-100, Enrich 70 SEC/ apparent CT663 (81) 7. 3 20 mM Tris, pH 7. 4 50 uL tetramer 0. 01% Triton X-100, 50 mM K-Na-phosphate, pH Enrich 70 SEC/ apparent CT663 (81) 50 mM K-Na- 7. 3 50 uL dimer phosphate, pH 7. 3 50 mM K-Na-phosphate, pH 1% PEG, 50 mM K- Enrich 70 SEC/ apparent CT663 (81) 7. 3 Na-phosphate, pH 7. 3 50 uL dimer

4. 3. 6 SPR analysis of Ct663/Scc1 binding to CopN

SPR was carried out at varies concentrations of ligand and analyte to optimize the experiment. The Scc1/CT663 ligand concentration was 4 nM. The serial analyte concentrations used were 2. 5, 5, 10, 20, and 40 nM of CopN. Raw data was collected in the form of a sensogram which plots response units (RU) vs time. The spectra corresponding to the blanks and the sample are automatically aligned for subtraction.

The data is first manually processed by selecting which blanks to subtract from the sample run. Sharp peaks, artifacts that occur during injection, are removed from the resulting spectrum. The BIAevaluation software was used to fit the data to a curve based on a 1:1 binding motif. The rate of association and dissociation of the

Scc1/CT663 complex and CopN is 1. 01 x 105 +/- 0. 02 x 105 1/Ms and 3. 04 x 10-4 +/-

0. 16 x 10-4 1/s respectively. The equilibrium dissociation constant is 3. 02 +/- 0. 17 nM, which suggest strong binding between CopN and the Scc1/CT663 complex.

Figure 4-8 The SPR sensogram corresponding to the binding event between the Scc1/CT663 complex and CopN. The calculated Kd=3. 02 +/- 0. 17 nM.

4. 3. 7 The magical CT663 sample from never never land

CT663 prepared by our collaborators was sent to us for HSQC analysis. The sample was said to have contained 500 µM CT663 without any purification tags and 20 mM sodium acetate at pH 4. 5. This sample is a good candidate for structural analysis because greater than 90% of the peaks are present. The peaks are well dispersed and of similar intensity, suggesting that the protein is fold and possibly in monomeric form.

This spectrum has not been reproducible because the exact conditions were not known.

There was a lack of communication from all parties involved in producing the sample and obtaining the spectrum. Recent meeting have revealed the appropriate conditions and attempts to reproduce this spectrum is underway.

Figure 4-9 1H-15N HSQC of 15N labeled CT663. Each peak corresponds to an 1H-15N bond from the backbone or a side chain of the protein.

4. 3. 8 NMR buffer optimization for CT663

NMR conditions are optimized to obtain good data. The buffer conditions and

protein concentration are varied until an appropriate spectrum is obtained. A good

HSQC has well resolve peaks of similar intensity, with one peak for every amino acid

except proline. Peak distribution is also important, as it gives information about the folding of the protein. Well disperse peaks are necessary for a structure determination.

1H-15N HSQC of 15N labeled CT663 in 20 mM sodium phosphate produced peaks that were dispersed. However, the peak intensities are not similar and not resolved between

8-8. 5 ppm. CT663 in 20 mM tris produce less peaks that vary in intensity, that lack dispersion. CT663 in 20 mM MES produced peaks that were not dispersed or similar in

90 intensity. Tris is a terrible buffer for CT663 because it does not allow for the detection of an adequate number of amide peaks. There are 142 amino acids that compose CT663,

accounting for the proline residues there should be 139 backbone amide peaks plus

additional side chain peaks. There are more peaks for samples in MES and phosphate

buffer; however, less than 90% of the amide peaks accounted. Peak dispersion is also a

problem, the peaks between 8-8. 5 ppm appear to be overlapping and of different

intensity. These peaks indicated that the protein may be aggregated or unfolded.

Figure 4-10 1H-15N HSQC of 15N labeled CT663 in 20 mM sodium phosphate, 20 mM tris, and 20 mM MES.

4. 3. 8 15N-CT663:Scc1 NMR analysis

The nature of the CT663/Scc1 complex formation is not known. In an effort to

analyze the effect of Scc1 on CT663, unlabeled Scc1 was added to 15N CT663 and the

1H-15N HSQC was analyzed. Figure 4-11 shows the overlapped 1H-15N HSQC for 15N

91 CT663 and 15N CT663 complexed with Scc1. The spectra show shifts in the amide

peaks that correspond to changes in the backbone structure. The inset of Figure 4-11

shows 13 peaks, 9 of which show varying degrees of shift, signifying a local structural

change.

Figure 4-11 Scc1 perturbs the amide peaks of 15N-CT663. An overlay of 1H-15N HSQC spectra of 15N-CT663 (red) and 15N-CT663:Scc1 (blue). The inset illustrates the selective shifts of the putative glycine amide peaks upon binding Scc1. Note: No peaks are covered by the inset.

4. 3. 9 NMR analysis of the CT663/Scc1 complex

Preliminary NMR data suggested that the Scc1/CT663 complex is a good sample

for structure determination, even though it is ~37 kDa. 15N-Scc1/CT663 was expressed and purified. The sample was given to Abigael Songok, a graduate student in our lab, for further analysis. The sample was desalted into 20 mM ammonium acetate at pH 5.

92 5, frozen at -80 oC, and lyophilized. The protein was reconstituted in 20 mM ammonium acetate and 10 mM DTT at pH 5. 5 with 10% D2O. The sample was concentrated to 250

µL and a 1H-15N HSQC spectrum was collected. The spectrum looks great with 95% of the amide peaks accounted. The inset of Figure 4-12 shows the peaks are dispersed in the zoomed view.

Figure 4-12 1H-15N HSQC of 15N labeled Scc1/CT663. The inset is a zoomed view to show the peak dispersion.

4. 4 Conclusions

The goal of this research is to analyze the structure and ligand interactions of

CT663. CT663 forms a tightly associated complex with Scc1 that cannot be disrupted with urea. This suggests that hydrophobic interactions play a minor role in keeping the

93 complex intact. A CopN pull down assay performed by Li Shen showed that CopN does not bind CT663 or Scc1 independently, but binds the Scc1/CT663 complex. What

remains unclear is where, spatially, CopN binds the complex. The Phyre2 server

predicted a linear structure, where CopN is bound mainly to CT663, and CT663 is

bound to Scc1. This quaternary structure is possible if the binding of CT663 to Scc1

causes a conformational change of CT663 that is favorable for CopN binding. It is also

possible that there is an area near the interface of the Scc1/CT663 complex where

CopN binds to both proteins. The binding kinetics of Scc1/CT663 was analyzed by SPR.

The interaction between the chaperone complex and CopN is strong bond with a Kd of

3. 02 +/- 0. 17 nM.

In order to obtain good data for structural analysis with NMR, proteins need to be

in their monomeric state or in small (< 25 kDa) multimeric state. The results of the gel

filtration experiments indicate that CT663 forms a dimer or a higher order oligomer

depending on the buffer conditions. Stabilizing the CT663 monomer is not

straightforward. The overlapped peaks in the center of the 1H-15N HSQC of 15N-CT663

in sodium phosphate and tris support the finding that CT663 is aggregating under these conditions. The 1H-15N HSQC of the 15N-CT663 sample sent to us by our collaborators

showed a very different scenario. The buffer conditions were never replicated because the protein we produced precipitated when the pH was lowered to 5. After several

meetings, it was discovered that the CT663 in the magical sample lacked the His-tag

used for purification. The His-tag increases the isoelectric point (pI) of CT663 causing it

to precipitate at pH 5 and may also be the source of the aggregation issues. The CT663

that is in complex with Scc1 does not have a His-tag, thus sarkosyl was used to

94 separate CT663 from the complex for NMR analysis. Abigael Songok is continuing this work. If the NMR spectra of CT663 cannot be replicated using sarkosyl to purify the protein, then CT663 without the His-tag can be produced and purified using a combination of chromatography methods.

Initial NMR experiments showed that when Scc1 is complexes CT663, there is a localized change in the backbone structure of CT663. In order to get a better understanding of how the structure of CT663 changes, the structure of the complex will be analyzed through complete and mixed isotopic labeling. The 1H-15N HSQC of 15N

Scc1/CT663 is very promising—it should lead to a full 3D structure of the complex.

Adding CopN to the isotopically labeled complex may give information, such as the residues of CT663 and Scc1 that bind CopN and changes to the CT663:Scc1 structure when CopN is bound. Knowledge of the binding, even between the complex and CopN, could lead to inhibitors that could be used in conjunction with current therapeutics to treat Chlamydia trachomatis.

95 References

1. Cooper, G. M. The cell: A molecular approach. The Molecular Composition of Cells 2000; 2nd edition:[Available from: http://www. ncbi. nlm. nih. gov/books/NBK9879/.

2. Nester, E. W. , Microbiology: A human perspective. 3rd Edition ed. 2001, Michigan: McGraw Hill Higher Education.

3. Lehninger, A. L. , D. L. Nelson, and M. M. Cox, Lehninger principles of biochemistry. 5th Edition ed. 2005, New York: W. H. Freeman.

4. Varki, A. and M. J. Chrispeels, Essentials of glycobiology. 1999, New York: Cold Spring Harbor Laboratory Press.

5. Banerjee, P. S. , G. W. Hart, and J. W. Cho, Chemical approaches to study o- glcnacylation. Chemical Society Reviews, 2013. 42(10): p. 4345-4357.

6. Santoyo-Ramos, P. , M. C. Castañeda-Patlán, and M. Robles-Flores, The role of o-linked beta-n-acetylglucosamine (glcnac) modification in cell signaling. 2012, Romania Intech.

7. Schleicher, E. D. and C. Weigert, Role of the hexosamine biosynthetic pathway in diabetic nephropathy. Kidney International Supplement, 2000. 58(S77): p. S13-S18.

8. Darley-Usmar, V. M. , L. E. Ball, and J. C. Chatham, Protein o-linked β-n- acetylglucosamine: A novel effector of cardiomyocyte metabolism and function. Journal of Molecular and Cellular Cardiology, 2012. 52(3): p. 538-549.

9. Fang, W. , Du, T. , Raimi, O. G. , Hurtado Guerrero, R. , Mariño, K. , Ibrahim, A. F. M. , Albarbarawi, O. , Ferguson, M. A. J. , Jin, C. and Van Aalten, D. M. F. , Genetic and structural validation of aspergillus fumigatus n- acetylphosphoglucosamine mutase as an antifungal target. Bioscience. Report, 2013. 33(5): p. 11.

10. Hart, G. W. , M. P. Housley, and C. Slawson, Cycling of o-linked beta-n- acetylglucosamine on nucleocytoplasmic proteins. Nature, 2007. 446(7139): p. 1017-22.

11. Whelan, S. and G. Hart, Identification of o-glcnac sites on proteins. Methods Enzymol, 2006. 415: p. 113 - 133.

12. Butkinaree, C. , K. Park, and G. Hart, O-linked n-acetylglucosamine (o-glcnac): Extensive crosstalk with phosphorylation to regulate signaling and transcription in

96 response to nutrients and stress. Biochem Biophys Acta, 2010. 2010(1800): p. 96 - 106.

13. Torres, C. and G. Hart, Topography and polypeptide distribution of terminal n- acetylglucosamine residues on the surfaces of intact lymphocytes. Evidence for o-linked glcnac. J Biol Chem, 1984. 259: p. 3308 - 3317.

14. Liu, Y. , X. Li, Y. Yu, J. Shi, Z. Liang, X. Run, Y. Li, C. -l. Dai, I. Grundke-Iqbal, K. Iqbal, F. Liu, and C. -X. Gong, Developmental regulation of protein o- glcnacylation, o-glcnac transferase, and o-glcnacase in mammalian brain. Plos One, 2012. 7(8): p. e43724.

15. Ma, J. and G. Hart, O-glcnac profiling: From proteins to proteomes. Clinical Proteomics, 2014. 11(1): p. 8.

16. Hanover, J. A. , M. W. Krause, and D. C. Love, The hexosamine signaling pathway: O-glcnac cycling in feast or famine. Biochimica et Biophysica Acta (BBA) - General Subjects, 2010. 1800(2): p. 80-95.

17. Khidekel, N. , S. Arndt, N. Lamarre-Vincent, A. Lippert, K. Poulin-Kerstien, B. Ramakrishnan, P. Qasba, and L. Hsieh-Wilson, A chemoenzymatic approach toward the rapid and sensitive detection of o-glcnac posttranslational modifications. Journal of the American Chemical Society, 2003. 125: p. 16162 - 16163.

18. Rexach, J. , P. Clark, D. Mason, R. Neve, E. Peters, and L. Hsieh-Wilson, Dynamic o-glcnac modification regulates creb-mediated gene expression and memory formation. Nature Chemical Biology, 2012. 8: p. 253 - 261.

19. Borman, S. Memories’ sweet origins. Available from: http://cen. acs. org/articles/90/web/2012/02/MemoriesSweet-Origins. html.

20. Rexach, J. E. , P. M. Clark, and L. C. Hsieh-Wilson, Chemical approaches to understanding o-glcnac glycosylation in the brain. Nature Chemical Biology, 2008. 4(2): p. 97-106.

21. Klein, S. B. and B. M. Thorne, Biological psychology. 2006, New York: Worth Publisher.

22. Yu, Y. , L. Zhang, X. Li, X. Run, Z. Liang, Y. Li, Y. Liu, M. H. Lee, I. Grundke- Iqbal, K. Iqbal, D. J. Vocadlo, F. Liu, and C. -X. Gong, Differential effects of an o- glcnacase inhibitor on tau phosphorylation. PLoS ONE, 2012. 7(4): p. e35277.

23. Yuzwa, S. A. , X. Shan, M. S. Macauley, T. Clark, Y. Skorobogatko, K. Vosseller, and D. J. Vocadlo, Increasing o-glcnac slows neurodegeneration and stabilizes tau against aggregation. Nature Chemical Biology, 2012. 8(4): p. 393-399.

97 24. Yuzwa, S. , A. Yadav, Y. Skorobogatko, T. Clark, K. Vosseller, and D. Vocadlo, Mapping o-glcnac modification sites on tau and generation of a site-specific o- glcnac tau antibody. Amino Acids, 2011. 40: p. 857 - 868.

25. Kamigaito, T. , T. Okaneya, M. Kawakubo, H. Shimojo, O. Nishizawa, and J. Nakayama, Overexpression of o-glcnac by prostate cancer cells is significantly associated with poor prognosis of patients. Prostate Cancer Prostatic Disease, 2014. 17(1): p. 18-22.

26. Gu, Y. , W. Mi, Y. Ge, H. Liu, Q. Fan, C. Han, J. Yang, F. Han, X. Lu, and W. Yu, Glcnacylation plays an essential role in breast cancer metastasis. Cancer Research, 2010. 70(15): p. 6344-6351.

27. Issad, T. , E. Masson, and P. Pagesy, O-glcnac modification, insulin signaling and diabetic complications. Diabetes & Metabolism, 2010. 36: p. 423 - 435.

28. Kahn, B. B. and J. S. Flier, Obesity and insulin resistance. The Journal of Clinical Investigation, 2000. 106(4): p. 473-481.

29. Konrad, R. J. , I. Mikolaenko, J. F. Tolar, K. Liu, and J. E. Kudlow, The potential mechanism of the diabetogenic action of streptozotocin: Inhibition of pancreatic beta-cell o-glcnac-selective n-acetyl-beta-d-glucosaminidase. Biochemical. Journal, 2001. 356(1): p. 31-41.

30. Lameira, J. , C. u. N. Alves, V. Moliner, S. Martí, R. Castillo, and I. a. Tuñón, Quantum mechanical/molecular mechanical molecular dynamics simulation of wild-type and seven mutants of cpnagj in complex with pugnac. The Journal of Physical Chemistry B, 2010. 114(20): p. 7029-7036.

31. Yuzwa, S. A. , M. S. Macauley, J. E. Heinonen, X. Shan, R. J. Dennis, Y. He, G. E. Whitworth, K. A. Stubbs, E. J. McEachern, G. J. Davies, and D. J. Vocadlo, A potent mechanism-inspired o-glcnacase inhibitor that blocks phosphorylation of tau in vivo. Nature Chemical Biology, 2008. 4(8): p. 483-490.

32. Dorfmueller, H. C. , V. S. Borodkin, M. Schimpl, and D. M. F. van Aalten, Glcnacstatins are nanomolar inhibitors of human o-glcnacase inducing cellular hyper-o-glcnacylation. Biochemical Journal, 2009. 420(2): p. 221-227.

33. Rexach, J. , C. Rogers, S. Yu, J. Tao, Y. Sun, and L. Hsieh-Wilson, Quantification of o-glycosylation stoichiometry and dynamics using resolvable mass tags. Nature Chemical Biology, 2010. 6: p. 645 - 651.

34. J T Gallagher, A. M. , and T M Dexter, Identification of two binding sites for wheat-germ agglutinin on polylactosamine-type oligosaccharides. Biochemical Journal, 1985. 231(1): p. 115-122.

35. Stalnaker, S. H. , S. Hashmi, J. -M. Lim, K. Aoki, M. Porterfield, G. Gutierrez- Sanchez, J. Wheeler, J. M. Ervasti, C. Bergmann, M. Tiemeyer, and L. Wells,

98 Site mapping and characterization of o-glycan structures on α-dystroglycan isolated from rabbit skeletal muscle. Journal of Biological Chemistry, 2010. 285(32): p. 24882-24891.

36. Wells, L. , K. Vosseller, R. N. Cole, J. M. Cronshaw, M. J. Matunis, and G. W. Hart, Mapping sites of o-glcnac modification using affinity tags for serine and threonine post-translational modifications. Molecular & Cellular Proteomics, 2002. 1(10): p. 791-804.

37. Sekine, O. , D. C. Love, D. S. Rubenstein, and J. A. Hanover, Blocking o-linked glcnac cycling in drosophila insulin-producing cells perturbs glucose-insulin homeostasis. Journal of Biological Chemistry, 2010. 285(49): p. 38684-38691.

38. Chung, S. , D. -O. Kang, J. Yamzon, D. Warburton, and C. J. Koh, O-glcnac mediated glycosylation down-regulation in mice with cyclophosphamide induced cystitis. The Journal of Urology, 2010. 183(1): p. 351-356.

39. Yee, A. , X. Chang, A. Pineda-Lucena, B. Wu, A. Semesi, B. Le, T. Ramelot, G. M. Lee, S. Bhattacharyya, P. Gutierrez, A. Denisov, C. -H. Lee, J. R. Cort, G. Kozlov, J. Liao, G. Finak, L. Chen, D. Wishart, W. Lee, L. P. McIntosh, K. Gehring, M. A. Kennedy, A. M. Edwards, and C. H. Arrowsmith, An nmr approach to structural proteomics. Proceedings of the National Academy of Sciences, 2002. 99(4): p. 1825-1830.

40. Higman, V. A. Protein nmr: A practical guide. 2012; Available from: http://www. protein-nmr. org. uk/.

41. Johannessen, M. , M. P. Delghandi, and U. Moens, What turns creb on? Cellular Signalling, 2004. 16(11): p. 1211-1227.

42. Zubenko, G. S. , H. B. Hughes, B. S. Maher, J. S. Stiffler, W. N. Zubenko, and M. L. Marazita, Genetic linkage of region containing the creb1 gene to depressive disorders in women from families with recurrent, early-onset, major depression. American Journal of Medical Genetics, 2002. 114(8): p. 980-987.

43. Gaiarsa, J. -L. , O. Caillard, and Y. Ben-Ari, Long-term plasticity at gabaergic and glycinergic synapses: Mechanisms and functional significance. Trends in Neurosciences. 25(11): p. 564-570.

44. Josselyn SA1, S. C. , Carlezon WA Jr, Neve RL, Nestler EJ, Davis M. , Long- term memory is facilitated by camp response element-binding protein overexpression in the amygdala. The Journal of Neuroscience, 2001. 21(7): p. 2404-2412.

45. Baba, T. , T. Ara, M. Hasegawa, Y. Takai, Y. Okumura, M. Baba, K. A. Datsenko, M. Tomita, B. L. Wanner, and H. Mori, Construction of escherichia coli k-12 in- frame, single-gene knockout mutants: The keio collection. Molecular Systems Biology, 2006. 2(16738554): p. 11.

99 46. Cormier, C. , J. Park, M. Fiacco, J. Steel, P. Hunter, J. Kramer, R. Singla, and J. LaBaer, Psi:Biology-materials repository: A biologist’s resource for protein expression plasmids. Journal of Structural and Functional Genomics, 2011. 12(2): p. 55-62.

47. Lazarus, B. D. , D. C. Love, and J. A. Hanover, Recombinant o-glcnac transferase isoforms: Identification of o-glcnacase, yes tyrosine kinase, and tau as isoform-specific substrates. Glycobiology, 2006. 16(16434389): p. 415-421.

48. Abràmoff, M. D. , P. J. Magalhães, and S. J. Ram, Image processing with imagej. Biophotonics international, 2004. 11(7): p. 36-43.

49. Frey, P. A. , F. C. Kokesh, and F. H. Westheimer, A reporter group at the active site of acetoacetate decarboxylase. I. Ionization constant of the nitrophenol. Journal of the American Chemical Society, 1971. 93(5127415): p. 7266-7269.

50. Projan, S. J. and R. P. Novick, Reciprocal intrapool variation in plasmid copy numbers: A characteristic of segregational incompatibility. Plasmid, 1984. 12(6494316): p. 52-60.

51. Lim KH, H. C. , Chang HI, Production of o-glcnac modified recombinant proteins in escherichia coli. Journal Of Microbiology And Biotechnology, 2002. 12.

52. Riu, I. -H. , I. -S. Shin, and S. -I. Do, Sp1 modulates ncogt activity to alter target recognition and enhanced thermotolerance in e. Coli. Biochemical and Biophysical Research Communications, 2008. 372(18486602): p. 203-209.

53. Lim, K. -H. and H. -I. Chang, O-linked n-acetylglucosamine suppresses thermal aggregation of sp1. FEBS Letters, 2006. 580(16879824): p. 4645-4652.

54. Liu G, H. Y. , Xiao R, Wang D, Acton TB, Montelione GT, Nmr structure of f-actin binding domain of arg/abl2 from homo sapiens. Proteins: Structure, Function, and Bioinformatics, 2010. 78(5): p. 1326–1330.

55. Bhonagiri, P. , G. R. Pattar, K. M. Habegger, A. M. McCarthy, L. Tackett, and J. S. Elmendorf, Evidence coupling increased hexosamine biosynthesis pathway activity to membrane cholesterol toxicity and cortical filamentous actin derangement contributing to cellular insulin resistance. Endocrinology, 2011. 152(21712361): p. 3373-3384.

56. Hedou, J. , B. Bastide, A. Page, J. Michalski, and W. Morelle, Mapping of o- linked beta-n-acetylglucosamine modification sites in key contractile proteins of rat skeletal muscle. Proteomics, 2009. 9: p. 2139 - 2148.

57. Gupta, R. and S. Brunak, Prediction of glycosylation across the human proteome and the correlation to protein function. Pacific Symposium On Biocomputing. , 2002. 7: p. 310 - 322.

100 58. Lubas, W. A. and J. A. Hanover, Functional expression of o-linked glcnac transferase. Domain structure and substrate specificity. Journal Biological Chemistry, 2000. 275(10753899): p. 10983-10988.

59. Yem, D. W. and H. C. Wu, Isolation of escherichia coli k-12 mutants with altered level of beta-n-acetylglucosaminidase. Journal of Bacteriology, 1976. 125(1107312): p. 372-373.

60. Cheng, Q. , H. Li, K. Merdek, and J. T. Park, Molecular characterization of the beta-n-acetylglucosaminidase of escherichia coli and its role in cell wall recycling. Journal of Bacteriology, 2000. 182(10940025): p. 4836-4840.

61. Votsch, W. and M. F. Templin, Characterization of a beta -n- acetylglucosaminidase of escherichia coli and elucidation of its role in muropeptide recycling and beta -lactamase induction. Journal Biological Chemistry, 2000. 275(10978324): p. 39032-39038.

62. Beer D, M. J. , Rast DM, Vasella A. ,, Synthesis of 2-acetamido-2-deoxy-d- gluconhydroximolactone- and chitobionhydroximolactone-derived n- phenylcarbamates, potential inhibitors of β-n-acetylglucosaminidase. Helvetica Chimica Acta, 1990. 73(7): p. 1918–1922.

63. Haltiwanger, R. S. , K. Grove, and G. A. Philipsberg, Modulation of o-linked n- acetylglucosamine levels on nuclear and cytoplasmic proteins in vivo using the peptide o-glcnac-beta-n-acetylglucosaminidase inhibitor o-(2-acetamido-2-deoxy- d-glucopyranosylidene)amino-n-phenylcarbamate. Journal Biological Chemistry, 1998. 273(9452489): p. 3611-3617.

64. Stubbs, K. A. , M. Balcewich, B. L. Mark, and D. J. Vocadlo, Small molecule inhibitors of a glycoside hydrolase attenuate inducible ampc-mediated beta- lactam resistance. Journal Biological Chemistry, 2007. 282(17439950): p. 21382- 21391.

65. Horsch, M. , L. Hoesch, A. Vasella, and D. M. Rast, N-acetylglucosaminono-1,5- lactone oxime and the corresponding (phenylcarbamoyl)oxime. Novel and potent inhibitors of beta-n-acetylglucosaminidase. European Journal of Biochemistry, 1991. 197(2029909): p. 815-818.

66. Deng, M. -D. , D. K. Severson, A. D. Grund, S. L. Wassink, R. P. Burlingame, A. Berry, J. A. Running, C. A. Kunesh, L. Song, T. A. Jerrell, and R. A. Rosson, Metabolic engineering of escherichia coli for industrial production of glucosamine and n-acetylglucosamine. Metabolic Engineering, 2005. 7(3): p. 201-214.

67. Rodríguez-Díaz, J. , A. Rubio-del-Campo, and M. J. Yebra, Metabolic engineering of lactobacillus casei for production of udp-n-acetylglucosamine. Biotechnology and Bioengineering, 2012. 109(7): p. 1704-1712.

101 68. Uehara, T. and J. T. Park, The n-acetyl-d-glucosamine kinase of escherichia coli and its role in murein recycling. Journal of Bacteriology, 2004. 186(21): p. 7273- 7279.

69. Cheng, Q. , H. Li, K. Merdek, and J. T. Park, Molecular characterization of the β- n-acetylglucosaminidase of escherichia coliand its role in cell wall recycling. Journal of Bacteriology, 2000. 182(17): p. 4836-4840.

70. Triassi, A. J. , M. S. Wheatley, M. Savka, H. M. Gan, R. C. J. Dobson, and A. O. B. Hudson, L,l-diaminopimelate aminotransferase (dapl): A putative target for the development of narrow-spectrum antibacterial compounds. Frontiers in Microbiology, 2014. 5.

71. Kreppel, L. K. and G. W. Hart, Regulation of a cytosolic and nuclear o-glcnac transferase: Role of the tetratricopeptide repeats. Journal of Biological Chemistry, 1999. 274(45): p. 32015-32022.

72. Baba, T. , T. Ara, M. Hasegawa, Y. Takai, Y. Okumura, M. Baba, K. A. Datsenko, M. Tomita, B. L. Wanner, and H. Mori, Construction of escherichia coli k‐12 in‐ frame, single‐gene knockout mutants: The keio collection. 2006.

73. Álvarez-Añorve, L. I. , M. L. Calcagno, and J. Plumbridge, Why does escherichia coli grow more slowly on glucosamine than on n-acetylglucosamine? Effects of enzyme levels and allosteric activation of glcn6p deaminase (nagb) on growth rates. Journal of Bacteriology, 2005. 187(9): p. 2974-2982.

74. Barton, L. , Structural and functional relationships in prokaryotes. 2005, New York: Springer.

75. Gurcel, C. , A. -S. Vercoutter-Edouart, C. Fonbonne, M. Mortuaire, A. Salvador, J. -C. Michalski, and J. Lemoine, Identification of new o-glcnac modified proteins using a click-chemistry-based tagging. Analytical and Bioanalytical Chemistry, 2008. 390(8): p. 2089-2097.

76. Brunham, R. C. and J. Rey-Ladino, Immunology of chlamydia infection: Implications for a chlamydia trachomatis vaccine. Nature Reviews Immunology, 2005. 5(2): p. 149-161.

77. CDC. Cdc fact sheet. 2013 [cited 2014 October 1, ]; Available from: http://www. cdc. gov/std/stats/sti-estimates-fact-sheet-feb-2013. pdf.

78. Paavonen, J. and W. Eggert-Kruse, Chlamydia trachomatis: Impact on human reproduction. Human Reproduction Update, 1999. 5(5): p. 433-447.

79. Somani, J. , V. B. Bhullar, K. A. Workowski, C. E. Farshy, and C. M. Black, Multiple drug-resistant chlamydia trachomatis associated with clinical treatment failure. Journal of Infectious Diseases, 2000. 181(4): p. 1421-1427.

102 80. Scott LaMontagne, D. , K. Baster, L. Emmett, T. Nichols, S. Randall, L. McLean, P. Meredith, V. Harindra, J. M. Tobin, G. S. Underhill, W. Graham Hewitt, J. Hopwood, T. Gleave, A. K. Ghosh, H. Mallinson, A. R. Davies, G. Hughes, and K. A. Fenton, Incidence and reinfection rates of genital chlamydial infection among women aged 16–24 years attending general practice, family planning and genitourinary medicine clinics in england: A prospective cohort study by the chlamydia recall study advisory group. Sexually Transmitted Infections, 2007. 83(4): p. 292-303.

81. Wenman, W. M. and R. U. Meuser, Chlamydia trachomatis elementary bodies possess proteins which bind to eucaryotic cell membranes. Journal of Bacteriology, 1986. 165(2): p. 602-607.

82. Bao, X. , B. E. Nickels, and H. Fan, Chlamydia trachomatis protein grga activates transcription by contacting the nonconserved region of sigma(66). Proceedings of the National Academy of Sciences of the United States of America, 2012. 109(42): p. 16870-16875.

83. Bloom, B. R. and P. H. Lambert, The vaccine book. 2003, California: Academic Press.

84. Archuleta, T. L. , Y. Du, C. A. English, S. Lory, C. Lesser, M. D. Ohi, R. Ohi, and B. W. Spiller, The chlamydia effector chlamydial outer protein n (copn) sequesters tubulin and prevents microtubule assembly. Journal of Biological Chemistry, 2011. 286(39): p. 33992-33998.

85. Pais, S. V. , C. Milho, F. Almeida, and L. J. Mota, Identification of novel type iii secretion chaperone-substrate complexes of chlamydia trachomatis. Plos One, 2013. 8(2).

86. Büttner D , a. H. S. Y. , Type iii protein secretion in plant pathogenic bacteria. Plant Physiology, 2009. 150(4): p. 1656-1664

87. Betts-Hampikian, H. J. and K. Fields, The chlamydial type iii secretion mechanism: Revealing cracks in a tough nut. Frontiers in Microbiology, 2010. 1.

88. Parsot, C. , C. Hamiaux, and A. L. Page, The various and varying roles of specific chaperones in type iii secretion systems. Current Opinion in Microbiology, 2003. 6(1): p. 7-14.

89. Rao, X. C. , P. Deighan, Z. Y. Hua, X. M. Hu, J. Wang, M. Luo, J. Wang, Y. M. Liang, G. M. Zhong, A. Hochschild, and L. Shen, A regulator from chlamydia trachomatis modulates the activity of rna polymerase through direct interaction with the beta subunit and the primary sigma subunit. Genes & Development, 2009. 23(15): p. 1818-1829.

103 90. Huang, J. , C. F. Lesser, and S. Lory, The essential role of the copn protein in chlamydia pneumoniae intracellular growth. Nature, 2008. 456(7218): p. 112- 115.

91. Santini, S. , S. Di Agostino, E. Coppari, A. R. Bizzarri, G. Blandino, and S. Cannistraro, Interaction of mutant p53 with p73: A surface plasmon resonance and atomic force spectroscopy study. Biochimica et Biophysica Acta (BBA) - General Subjects, 2014. 1840(6): p. 1958-1964.

92. Izore, T. , V. Job, and A. Dessen, Biogenesis, regulation, and targeting of the type iii secretion system. Structure, 2011. 19(5): p. 603-612.

93. Krieger, E. , S. B. Nabuurs, and G. Vriend, Homology modeling, in Structural bioinformatics2005, John Wiley & Sons, Inc. p. 509-523.

94. Cooper, C. , Zhang, K. , Andres, S. N. , Hannemann, M. , Junop, M. S. , Coombes, B. , Structural analysis and quantitative proteomic interactome of a novel virulence chaperone in salmonella. Protein Data Bank, 2009.

95. Galán, J. E. and A. Collmer, Type iii secretion machines: Bacterial devices for protein delivery into host cells. Science, 1999. 284(5418): p. 1322-1328.

96. Phan, J. , J. E. Tropea, and D. S. Waugh, Structure of the yersinia pestis type iii secretion chaperone sych in complex with a stable fragment of yscm2. Acta Crystallographica Section D, 2004. 60(9): p. 1591-1599.

97. Luo, Y. , M. G. Bertero, E. A. Frey, R. A. Pfuetzner, M. R. Wenk, L. Creagh, S. L. Marcus, D. Lim, F. Sicheri, C. Kay, C. Haynes, B. B. Finlay, and N. C. J. Strynadka, Structural and biochemical characterization of the type iii secretion chaperones cest and sige. Nature Structural & Molecular Biology, 2001. 8(12): p. 1031-1036.

98. Stebbins, C. E. and J. E. Galan, Maintenance of an unfolded polypeptide by a cognate chaperone in bacterial type iii secretion. Nature, 2001. 414(6859): p. 77- 81.

99. Schubot, F. D. , M. W. Jackson, K. J. Penrose, S. Cherry, J. E. Tropea, G. V. Plano, and D. S. Waugh, Three-dimensional structure of a macromolecular assembly that regulates type iii secretion in yersinia pestis. Journal of Molecular Biology, 2005. 346(4): p. 1147-1161.

100. Deane, J. E. , P. Roversi, C. King, S. Johnson, and S. M. Lea, Structures of the shigella flexneri type 3 secretion system protein mxic reveal conformational variability amongst homologues. Journal of Molecular Biology, 2008. 377(4): p. 985-992.

104 101. Bennion, B. J. and V. Daggett, The molecular basis for the chemical denaturation of proteins by urea. Proceedings of the National Academy of Sciences, 2003. 100(9): p. 5142-5147.

105 Appendix

Letter of Permission

This is the permission to use my publication in chapter 2 of this thesis.

106

Supplementary Figures

Supplementary Figure 1. Ni-affinity purified, resolubilized pellet from the expression of CREB1-Q2-bZIP (pET-28b-CREB1-Q2-bZIP) without (lane1) and with (lane2) co- expression of OGT (pET-43. 1-OGT) in E. coli BL21-Gold(DE3) analyzed with (a) Coomassie-stained SDS-PAGE and (b) anti-O-GlcNAc MAb CTD 110. 6 western blot. HEWL, ovalbumin, and protein ladder are shown in the last three lanes, respectively. HEWL acts as a positive control for the western blot because the secondary antibody (DyLight 550-conjugate Goat anti-Mouse IgG) binds non-specifically to HEWL. Residual HEWL can be seen in lanes 1 and 2 because it was used to lyse the cells in these preparations. HEWL was not used to lyse cells in any of the other experiments presented in this manuscript because of its interference in western blot analysis. The western blot of Q2-bZIP (lanes 1 and 2) demonstrates that OGT is modifying the protein with O-GlcNAc, consistent with the analysis of Q2-bZIP in the lectin blot in Figure 2. The pellets were resolubilized in 6 M urea, centrifuged at 14,000 g at 4 °C for 30 min. Q2- bZIP was purified from the supernatant using a Bio-Rad BioLogic DuoFlow chromatography system with a 1 mL His-Select Ni-affinity gel column. After loading the supernatant, the column was washed with 10 mL of 6 M urea in buffer (50 mM sodium phosphate, 300 mM NaCl, pH 8. 0), 30 mL gradient (6 – 0 M urea in buffer) at a flow rate of 0. 1 mL/min, 10 mL of buffer, and 10 mL of 300 mM imidazole in buffer.

107

Supplementary Figure 2. The CXC fractions of cleared lysate from a culture of BL21- Gold(DE3) cells expressing empty pET-28b were analyzed using the MU-GlcNAc well- plate assay with the β-N-acetylglucosaminidase negative (-) and positive (+) controls. The active CXC fractions (#4-16) were combined and used to determine the enzyme kinetic parameters and activity towards the glycopeptide. Error bars represent the standard deviation of three replicate samples. The positive control intensity was truncated to better visualize the sample data.

Supplementary Figure 3. Velocities of pNP-GlcNAc hydrolysis by E. coli β-N- acetylglucosaminidase, partially purified with CXC from BL21-Gold(DE3) cleared lysate, at varying pNP-GlcNAc concentrations. The velocities (nmoles of nitrophenolate formed per mL of fraction/min) were calculated from the absorbance at 405 nm of samples measured after 4, 9, and 14 min of reaction time. The curve was fitted to the Michaelis- Menten equation (1) using nonlinear regression analysis to give a Km value of 0. 35 mM.

108

Supplementary Figure 4. Full 1H NMR spectra of samples from testing the E. coli β-N- acetylglucosaminidase activity of CXC fractions towards the synthetic O-GlcNAc- modified peptide, SVES(β-O-GlcNAc)GSADAK-NH2. Each spectrum was treated identically for direct comparison.

109

Supplementary Figure 5. Cleared lysate from cultures of (a) ∆nagZ K12 strain and (b) parent K12 strain were separated with CXC and analyzed with the MU-GlcNAc well- plate assay with the β-N-acetylglucosaminidase positive control (+) and negative control (-). Error bars represent the standard deviation of three replicate samples. The positive control intensity was truncated to better visualize the sample data.

110

Supplementary Figure 6. Full 1H NMR spectra of samples from testing for E. coli β-N- acetylglucosaminidase activity in ∆nagZ K12 lysate towards the synthetic O-GlcNAc- modified peptide, SVES(β-O-GlcNAc)GSADAK-NH2. Each spectrum was treated identically for direct comparison.

111 Vita Octavia Y Goodwin earned her B. S. in chemistry from Georgia Southern

University in May of 2008. Octavia started her career as a research scientist in 2006.

Under the supervision of Dr. Karen Welch, Octavia synthesized azo dyes. She also gained some computational molecular modeling experience while working on a glycomemtic ligand docking project. In January of 2009 Octavia joined Dr. Megan

Macnaughtan’s research group. While working for Dr. Macnaughtan, she bio- engineered a system to express near milligram quantities of O-GlcNAc-modified protein for structural analysis. Octavia will be receiving a Doctor of Philosophy in chemistry from Louisiana State University in December of 2014.

112